Methods of system analysis. System analysis of foreign trade relations of the agro-industrial complex of the region Methodological approaches and methods of system analysis

System analysis involves: the development of a systematic method for solving a problem, i.e. a logically and procedurally organized sequence of operations aimed at choosing the preferred alternative for solving a problem. System analysis is implemented practically in several stages, however, there is still no unity regarding their number and content, because. There is a wide variety of applied problems in science.

In progress system analysis different methods are used at different levels. At the same time, the system analysis itself plays the role of the so-called. a methodological framework that combines all the necessary methods, research techniques, activities and resources to solve problems. In essence, systems analysis organizes our knowledge of a problem in such a way as to help select the appropriate strategy for solving it or predict the results of one or more strategies that seem appropriate to those who must make decisions to resolve the contradiction that gave rise to the problem. In the most favorable cases, the strategy found through systems analysis is "best" in some specific sense.

Consider system analysis methodology on the example of the theory of the English scientist J. Jeffers, which suggests highlighting seven stages .

Stage 1 "Problem selection". The realization that there is some problem that can be investigated with the help of systems analysis, important enough to study in detail. The very understanding that a truly systematic analysis of the problem is needed is as important as choosing the right research method. On the one hand, one can tackle a problem that is not amenable to system analysis, and on the other hand, one can choose a problem that does not require the full power of system analysis for its solution, and it would be uneconomical to study by this method. This duality of the first stage makes it critical to the success or failure of the entire study.

Stage 2 "Statement of the problem and limitation of its complexity." Once the existence of the problem is recognized, it is required to simplify the problem so that it is likely to have an analytical solution, while retaining all those elements that make the problem interesting enough for practical study. Here again we are dealing with a critical stage in any systems research. It is at this stage that you can make the most significant contribution to solving the problem. The success or failure of the whole study depends largely on a delicate balance between simplification and complexity - a balance that retains all the links to the original problem that are sufficient for the analytical solution to be interpretable. The problem may not be solved due to the fact that the accepted level of complexity will make it difficult for subsequent modeling, not allowing to obtain its solution.

Stage 3 "Establishing a hierarchy of goals and objectives." After setting the task and limiting the degree of its complexity, you can begin to set the goals and objectives of the study. Usually these goals and objectives form a certain hierarchy, with the main tasks being successively subdivided into a number of secondary ones. In such a hierarchy, it is necessary to prioritize the various stages and correlate them with the efforts that need to be made to achieve the goals set. Thus, in a complex study, it is possible to assign relatively low priority to those goals and objectives that, although important from the point of view of obtaining scientific information, have a rather weak influence on the type of decisions made regarding the impact on the system and its management. In another situation, when this task is part of the program of some fundamental research, the researcher is deliberately limited to certain forms of management and concentrates maximum efforts on tasks that are directly related to the processes themselves. In any case, for the fruitful application of systems analysis, it is very important that the priorities assigned to the various tasks are clearly defined.

Stage 4 "Choosing ways to solve problems." At this stage, the researcher can usually choose several ways to solve the problem. As a rule, families of possible solutions to specific problems are immediately visible to an experienced systems analyst. Each specific problem can usually be solved in more than one way. Again, the choice of the family within which to search for an analytical solution depends on the experience of the systems analyst. An inexperienced researcher can spend a lot of time and money trying to apply a solution from any family, not realizing that this solution was obtained under assumptions that are unfair for the particular case with which he is dealing. The analyst, on the other hand, often develops several alternative solutions and only later settles on the one that best suits his task.

Stage 5 "Modeling". Once suitable alternatives have been analyzed, the next important step is to model the complex dynamic relationships between different aspects of the problem. At the same time, it should be remembered that the processes being modeled, as well as the feedback mechanisms, are characterized by internal uncertainty, and this can significantly complicate both the understanding of the system and its controllability. In addition, the modeling process itself must take into account a complex set of rules that will need to be observed when deciding on an appropriate strategy. At this stage, it is very easy to get carried away by the elegance of the model, and as a result, all points of contact between the real decision-making processes and the mathematical apparatus will be lost. In addition, when developing a model, unverified hypotheses are often included in it, and it is rather difficult to predetermine the optimal number of subsystems. It can be assumed that a more complex model better takes into account the complexities of a real system, but although this assumption seems to be intuitively correct, additional factors must be taken into account. Consider, for example, the hypothesis that a more complex model also gives higher accuracy in terms of the uncertainty inherent in model predictions. Generally speaking, the systematic bias that occurs when a system is decomposed into several subsystems is inversely related to the complexity of the model, but there is also a corresponding increase in uncertainty due to errors in measuring individual model parameters. Those new parameters that are introduced into the model must be quantified in field and laboratory experiments, and there are always some errors in their estimates. After going through the simulation, these measurement errors contribute to the uncertainty of the resulting predictions. For all these reasons, in any model it is advantageous to reduce the number of subsystems included in the consideration.

Stage 6 "Assessment of possible strategies". Once the simulation has been brought to the stage where the model can be used, the stage of evaluating the potential strategies derived from the model begins. If it turns out that the underlying assumptions are incorrect, you may have to return to the modeling stage, but it is often possible to improve the model by slightly modifying the original version. It is usually also necessary to investigate the “sensitivity” of the model to those aspects of the problem that were excluded from the formal analysis at the second stage, i.e. when the task was set and the degree of its complexity was limited.

Stage 7 "Implementation of results". The final stage of the system analysis is the practical application of the results obtained in the previous stages. If the study was carried out according to the above scheme, then the steps that need to be taken for this will be quite obvious. However, systems analysis cannot be considered complete until the research reaches the stage of practical application, and it is in this respect that much of the work done has been left unfulfilled. At the same time, just at the last stage, the incompleteness of certain stages or the need to revise them may be revealed, as a result of which it will be necessary to go through some of the already completed stages again.

Thus, the purpose of multi-stage systems analysis is to help choose the right strategy for solving practical problems. The structure of this analysis is intended to focus the main effort on complex and usually large-scale problems that cannot be solved by simpler methods of research, such as observation and direct experimentation.

Levels of decision making on a problem. The process of developing and making decisions on a problem can be represented as a set of methods and techniques of activity of a decision maker (DM). At the same time, the decision maker is guided by certain provisions, guidelines, principles, striving to organize the most effective system that will allow developing the optimal solution in a given situation. In this process, based on the decision-making mechanism, it is possible to single out separate levels, the elements of which the decision maker invariably encounters.

The main levels of decision-making on the problem:

1. Individual-semantic level. Decision-making at this level is carried out by the decision maker on the basis of logical reasoning. At the same time, the decision-making process depends on the individual experience of the decision maker and is closely related to the change in the specific situation. Based on this, people at the semantic level cannot understand each other, and the decisions they make are often not only unreasonable, but also devoid of organizational meaning. Thus, at this level, decisions are made only on the basis of "common sense".

2. Communicative-semantic level. At this level, decisions are already made on the basis of the communicative interaction of the persons involved in the decision-making. Here we are not talking about traditional communication, but about specially selected communication. The organizer of communication - the decision maker "launches" communication when there is a difficulty in the activity that gives rise to a problem situation. Participants in communication in the same situation can see different things based on their subjective position. As a result, the decision maker personally or with the help of an arbitrator organizes justified criticism and arbitration evaluation of various points of view. At this level, there is a merging of individual points of view with generally valid ones.

The first and second levels are considered pre-conceptual. It is at these levels that the leaders of organizations most often make decisions.

3. Conceptual level. At this level, there is a departure from individual opinions, and strict concepts are used. This stage involves the use of special tools for professional communication of decision makers with interested specialists, which helps to improve the quality of their professional interaction in the process of developing a solution.

4. problematic level. At this level, in order to solve problems, it is necessary to move from an individual semantic understanding of the problem situation that has developed in the decision-making process, to understanding it through meanings. If the goal of the decision maker is to solve a specific problem, known algorithms are used and the development of simple procedures is required. When the decision maker is faced with a certain problem and there is a situation of uncertainty, the decision is made by building a theoretical model, formulating hypotheses, developing solutions using a creative approach. Difficulties in this activity should lead to the next level of decision-making - systemic.

5. System level. This level requires the decision maker to have a systematic vision of all elements of the decision-making environment, the integrity of the representation of the control object and the interaction of its parts. Interaction should be transformed into mutual assistance of elements of integrity, which provides a systemic effect from the activity.

6. Universal-system level. Making a decision at this level involves the decision maker's vision of integrity in the control object and its integration into the environment. Empirical observations and the resulting analytical information are used here to determine the development trends of the object. The level requires the decision maker to build a complete picture of the surrounding world.

Thus, it is difficult for decision makers to move from level to level in making a decision on the problem. This may be his subjective doubts or the objective need to solve problems and problems, taking into account the requirements of a particular level. The more complex the control object (problem), the more high level decision making is required. At the same time, a certain decision-making mechanism must correspond to each level, it is also necessary to use level criteria for choosing a course of action.

Comparison of intuitive and systematic approach to decision making on a problem. In a situation where we need to make some decision on a problem (we assume that we make this decision on our own, in other words, it is not “imposed on us”), then we can act to determine which particular decision is better to take. two fundamentally different methods.

First method is simple and operates entirely on the basis of previously acquired experience and acquired knowledge. Briefly, it is as follows: having in our mind the initial situation, we

1) we select in memory one or several patterns known to us (“template”, “system”, “structure”, “principle”, “model”) that have a satisfactory (in our opinion) analogy with the initial situation;

2) we apply for the current situation a solution that corresponds to the best solution for an already known pattern, which in this situation becomes a model for its adoption.

This process of mental activity occurs, as a rule, unconsciously, and this is the reason for its extraordinary effectiveness. Due to our “unconsciousness”, we will call this decision-making method “intuitive”. However, it should be noted that this is nothing more than a practical application of one's previous experience and acquired knowledge. Do not confuse intuitive decision making with fortune telling or coin tossing. intuition in this case there is an unconscious quintessence of knowledge and experience of the person making the decision. Therefore, intuitive solutions are often very successful, especially if the person has sufficient experience in solving similar problems.

Second method is much more complicated and requires the involvement of conscious mental efforts aimed at applying the method itself. Briefly describe it as follows: having in our mind the initial situation, we

1) we select some efficiency criterion to evaluate the future solution;

2) determine the reasonable boundaries of the system under consideration;

3) we create a system model suitable for analogy with the initial situation;

4) explore the properties and behavior of this model to find the best solution;

5) apply the found solution in practice.

This complex decision-making method, as we already know, is called "systemic" due to the conscious application of the concepts of "system" and "model". The key in it is the task of competent development and use of models, because it is the model that is the result we need, which, moreover, can be remembered and used repeatedly in the future for similar situations.

If we compare these two methods with each other, then at first glance the effectiveness of the "intuitive" approach is obvious both in terms of the speed of decision-making and the cost of the efforts made. And indeed it is.

And what is the advantage of the "systemic" method, if any?

The fact is that the intuitive approach gives us an initially known solution to the task or problem situation, and using a systematic approach, we really do not know the solution we are looking for until some point. And this means that the practice of a systematic approach is "inherent" in people by nature and is to the same extent the basis of a person's personal training (especially clearly in his first years of life).

Intuitive and systematic decision-making methods do not contradict each other. However, each of them is more appropriate to use in a situation that is suitable for him. To find out in what situations what is better to use, let's first consider the following illustrative example.

Example. Let's imagine a situation when you enter the building of the Institute. To enter you must open and go through the entrance door. You have done this many times already, and, of course, you don’t think about it, that is, you do it “automatically”. Although, if you look at it, these actions are a rather complex coordinated chain of movements of the arms, legs and body of the body: not a single robot modern development technology and the success of artificial intelligence is not yet able to do this as naturally as, however, and just walk too. However, you do it easily and freely, because there are already well-functioning specific behaviors in the spinal cord and lower brain that give the correct result of predictions of your actions to open the door without using the resources of higher brain regions for this task. In other words, in such cases we use an already established decision-making model.

Now let's assume that the spring was replaced while you were away and that much more force is needed to open it. What will happen? As usual, you approach, take the handle, press ..., but the door does not open. If at this moment you are in thought, then you can even unsuccessfully pull the door handle several times until your nervous system will not get through to the consciousness that the situation requires study and some special reaction. What happened? The old model, which previously worked flawlessly for this situation, did not work - the prediction did not give the expected result. Therefore, you study what happened now, find the cause of the problem, understand that you need to make more significant efforts to open the door and determine what specific efforts. Then you “automatically update the model” of behavior for this situation and soon enough, probably within one day, the new model will “take root” and then you, as before, will enter your institute without thinking about it.

In this case, we took a "systemic" approach - we examined the situation, changed the unusable model and "put it into operation."

This simple example shows how our organism effectively applies modeling in practice in a systematic approach to making a decision on a problem. This combination is the reason for the extremely high ability of a person to adapt to new and unfavorable conditions. In a situation of uncertainty, when old models do not work, we develop and apply new ones, which should then work well for similar situations. This is the effect of learning, or rather the acquisition of a skill.

REMEMBER: Approaching the solution of fundamentally new tasks, we must immediately apply a systematic approach, spend additional efforts on its implementation, and not wait for inevitable problems with the implementation of the project.

The practice of applying a systematic approach when making a decision on a problem in most cases does not require serious involvement of expensive resources, the use of special software and a complete description of any processes. It happens that one brainstorming session, sheets of paper and a pencil with an eraser are enough to successfully solve a specific problem.

So, a systematic approach to decision-making on a problem involves following a clear algorithm consisting of 6 steps:

· problem definition;

· determination of criteria for choosing a solution;

· assigning weights to criteria;

· development of alternatives;

· evaluation of alternatives;

· choosing the best alternative.

However, there are circumstances such as: high level of uncertainty, lack or insufficiency of precedents, limited facts, evidence that points ambiguously the right way, analytical data of little usability, few good alternatives, limited time does not always allow for a systematic approach.

In this case, the decision maker is required to show creativity- i.e. the solution must be creative, original, unexpected. creative solution is born in the presence of the following factors:

· the person making the decision must have relevant knowledge and experience;

· he must have creative abilities;

· work on decision-making should be supported by appropriate motivation.

Finally, the process of making a decision on the problem and the subsequent reaction to it is influenced by cognitive biases And organizational constraints.

cognitive biases can be categorized according to the decision-making stage at which these prejudices influence.

At the stage of information gathering:

availability of information- only easily accessible information is selected for problem analysis;

confirmation bias- from the entire array of information, only that one is selected for analysis that confirms the initial (conscious or subconscious) attitude of the decision maker.

At the stage of information processing:

· risk avoidance- the tendency to avoid risk at all costs, even in the face of a highly probable positive outcome if a moderate risk is taken;

· excessive confidence in someone or something;

· framing- the influence of the format or wording of the question on the answer to this question;

· anchoring- the tendency to rely excessively on single data when making a decision;

· (un)representativeness of the sample.

At the decision stage:

· bounded rationality- the tendency of a person, when mentally sorting through possible solutions, to stop at the first “tolerable” solution that comes across, ignoring the remaining options (among which, perhaps, there is a “best” solution);

· groupthink- the influence of the general position of a group of people on the individual position of a person;

· herd feeling;

· social norms;

· impression management- the process by which a person tries to control the impression made on other people;

· competitive pressure;

· ownership effect- a person tends to value more what he directly owns.

At the stage of reaction to the decision made:

· illusion of control- the conviction of a person in his control over the situation to a greater extent than it really is;

· forcing conviction- a situation in which a person continues to take action in support of the original decision (to prove the correctness of this decision) even after the error of the original decision has become apparent;

· judgment in hindsight- the tendency to judge the events that have come as if in the past they were easy to predict and reasonably expected;

· fundamental attribution error- the tendency of a person to explain successes by his personal merits, and failures - by external factors;

· subjective assessment- the tendency to interpret data in accordance with one's beliefs/preferences.

Organizational restrictions, such as the system of personnel assessment, the system of rewards and motivation, the formal regulation adopted in the organization, the established time limits and historical precedents for solving similar problems also affect the decision-making process.

Thus, a systematic approach makes it possible to identify new characteristics of the problem under study, and to build a model of its solution that is fundamentally different from the previous one.

conclusions

1. Any scientific, research and practical activity is carried out on the basis of methods (techniques or methods of action), methods (a set of methods and techniques for carrying out any work) and methodologies (a set of methods, rules for the distribution and assignment of methods, as well as work steps and their sequences). System analysis is a set of methods and tools for developing, adopting and justifying the optimal decision from many possible alternatives. It is used primarily to solve strategic problems. The main contribution of system analysis to the solution of various problems is due to the fact that it makes it possible to identify those factors and relationships that may later turn out to be very significant, that it makes it possible to change the observation technique and experiment in such a way as to include these factors in consideration, and highlights the weak points of hypotheses. and assumptions.

2. When applying systems analysis, the emphasis is on testing hypotheses through experiments and rigorous sampling procedures creates powerful tools for understanding the physical world and combines these tools into a system of flexible but rigorous study of complex phenomena. This method is considered as a methodology for in-depth understanding (understanding) and ordering (structuring) of the problem. Hence, the methodology of system analysis is a set of principles, approaches, concepts and specific methods, as well as techniques. In systems analysis, the emphasis is on developing new principles of scientific thinking that take into account the interconnection of the whole and contradictory trends.

3. System analysis is not something fundamentally new in the study of the surrounding world and its problems - it is based on a natural science approach. In contrast to the traditional approach, in which the problem is solved in a strict sequence of the above steps (or in a different order), the systems approach consists in the multiple-connectedness of the solution process. The main and most valuable result of system analysis is not a quantitatively defined solution to the problem, but an increase in the degree of its understanding and possible solutions among specialists and experts participating in the study of the problem, and, most importantly, among responsible persons who are provided with a set of well-developed and evaluated alternatives.

4. The most general concept that denotes all possible manifestations of systems is “systematic”, which is proposed to be considered in three aspects:

a) systems theory provides rigorous scientific knowledge about the world of systems and explains the origin, structure, functioning and development of systems of various nature;

b) a systematic approach - performs orientation and worldview functions, provides not only a vision of the world, but also orientation in it. The main feature of a systematic approach is the presence of a dominant role of a complex, not simple, whole, and not constituent elements. If, with the traditional approach to research, thought moves from the simple to the complex, from parts to the whole, from elements to the system, then with the systematic approach, on the contrary, thought moves from the complex to the simple, from the whole to its constituent parts, from the system to the elements. ;

c) system method - implements cognitive and methodological functions.

5. Systematic consideration of the object involves: the definition and study of systemic quality; identification of the totality of elements forming the system; establishing links between these elements; study of the properties of the environment surrounding the system, important for the functioning of the system, at the macro and micro levels; revealing the relationships connecting the system with the environment.

The system analysis algorithm is based on the construction of a generalized model that reflects all the factors and relationships of the problem situation that may appear in the solution process. The system analysis procedure consists in checking the consequences of each of the possible alternative solutions for choosing the optimal one according to any criterion or their combination.

Bertalanfi L. background. General systems theory - a review of problems and results. System Research: Yearbook. M.: Nauka, 1969. S. 30-54.

Boulding K. General systems theory - the skeleton of science // Studies in general systems theory. M.: Progress, 1969. S. 106-124.

Volkova V.N., Denisov A.A. Fundamentals of control theory and system analysis. SPb.: SPbGTU, 1997.

Hegel G.W.F. Science of logic. In 3 vols. M.: 1970 - 1972.

Dolgushev N.V. Introduction to applied systems analysis. M., 2011.

Dulepov V.I., Leskova O.A., Maiorov I.S. System ecology. Vladivostok: VGUEiS, 2011.

Zhivitskaya E.N. System analysis and design. M., 2005.

Kaziev V.M. Introduction to the analysis, synthesis and modeling of systems: lecture notes. M.: IUIT, 2003.

Kachala V.V. Fundamentals of system analysis. Murmansk: MSTU, 2004.

When an intuitive method is used, and when a system method of decision making is used. Rb.ru Business Network, 2011.

Concepts of modern natural science: lecture notes. M., 2002.

Lapygin Yu.N. Theory of organizations: textbook. allowance. M., 2006.

Nikanorov S.P. Systems Analysis: A Stage in the Development of Problem Solving Methodology in the United States (translation). M., 2002.

Fundamentals of system analysis. Working programm. St. Petersburg: SZGZTU, 2003.

Peregudov F.I., Tarasenko F.P. Introduction to system analysis. M.: high school, 1989.

Pribylov I. Decision making process/www.pribylov.ru.

Sadovsky V.N. System approach and general systems theory: status, main problems and development prospects. Moscow: Nauka, 1980.

Svetlov N.M. Theory of systems and system analysis. UMK. M., 2011.

CERTICOM - Management consulting. Kyiv, 2010.

System analysis and decision making: Dictionary-reference book / ed. V.N. Volkova, V.N. Kozlov. Moscow: Higher school, 2004.

System analysis: lecture notes. Website for methodological support of the system of information and analytical support for decision-making in the field of education, 2008.

Spitsnadel VN Osnovy sistemnogo analiza: ucheb. allowance. St. Petersburg: "Publishing House" Business Press ", 2000.

Surmin Yu.P. Theory of systems and system analysis: textbook. allowance. Kyiv: MLUP, 2003.

Theory of organization: textbook. allowance /partnerstvo.ru.

Fadina L.Yu., Shchetinina E.D. Management decision-making technology. Sat. NPC articles. M., 2009.

Khasyanov A.F. System analysis: lecture notes. M., 2005.

Chernyakhovskaya L.R. Systems methodology and decision making. Brief summary lectures. Ufa: UGATU, 2007.

Chepurnykh E.M. System analysis in the theory of state and law. Virtual club of lawyers/ http://www.yurclub.ru/docs/theory/article9.html.

The central procedure in system analysis is the construction of a generalized model (or models) that reflects all the factors and relationships of the real situation that may appear in the process of implementing the decision. The resulting model is investigated in order to find out the closeness of the result of applying one or another of the alternative options actions to the desired, the comparative cost of resources for each of the options, the degree of sensitivity of the model to various undesirable external influences. Systems analysis is based on a number of applied mathematical disciplines and methods widely used in modern activities management: operations research, peer review method, critical path method, queuing theory, etc. Technical basis system analysis -- modern computers and information systems.

The methodological means used in solving problems with the help of system analysis are determined depending on whether a single goal or a certain set of goals is pursued, whether one person or several people make a decision, etc. When there is one fairly clearly defined goal, the degree of achievement of which can be evaluated on the basis of one criterion, methods of mathematical programming are used. If the degree of achievement of the goal must be assessed on the basis of several criteria, the apparatus of utility theory is used, with the help of which the criteria are ordered and the importance of each of them is determined. When the development of events is determined by the interaction of several persons or systems, each of which pursues its own goals and makes its own decisions, the methods of game theory are used.

The effectiveness of the study of control systems is largely determined by the chosen and used research methods. To facilitate the choice of methods in real decision-making conditions, it is necessary to divide the methods into groups, characterize the features of these groups and give recommendations on their use in the development of models and methods of system analysis.

The whole set of research methods can be divided into three large groups: methods based on the use of knowledge and intuition of specialists; methods of formalized representation of control systems (methods of formal modeling of the processes under study) and integrated methods.

As already noted, a specific feature of system analysis is the combination of qualitative and formal methods. This combination forms the basis of any technique used. Let's consider the main methods aimed at using the intuition and experience of specialists, as well as methods of formalized representation of systems.

Methods based on the identification and generalization of the opinions of experienced experts, the use of their experience and non-traditional approaches to the analysis of the organization's activities include: the "Brainstorming" method, the "scenarios" type method, the method of expert assessments (including SWOT analysis), the " Delphi", methods such as "tree of goals", "business game", morphological methods and a number of other methods.

The above terms characterize one or another approach to enhancing the identification and generalization of the opinions of experienced experts (the term "expert" in Latin means "experienced"). Sometimes all these methods are called "expert". However, there is also a special class of methods that are directly related to the questioning of experts, the so-called method of expert assessments (since it is customary to put down marks in points and ranks in polls), therefore, these and similar approaches are sometimes combined with the term "qualitative" (specifying the convention of this name, since when processing the opinions received from specialists, quantitative methods can also be used). This term (although somewhat cumbersome) more than others reflects the essence of the methods that specialists are forced to resort to when they not only cannot immediately describe the problem under consideration by analytical dependencies, but also do not see which of the methods of formalized representation of systems considered above could help get the model.

Brainstorming methods. The concept of brainstorming has become widespread since the early 1950s as a "method of systematically training creative thinking" aimed at "discovering new ideas and reaching agreement among a group of people based on intuitive thinking."

Methods of this type pursue the main goal - the search for new ideas, their broad discussion and constructive criticism. The main hypothesis is the assumption that among a large number of ideas there are at least a few good ones. Depending on the rules adopted and the rigidity of their implementation, there are direct brainstorming, the method of exchange of opinions, methods such as commissions, courts (when one group makes as many proposals as possible, and the second tries to criticize them as much as possible), etc. Recently, sometimes brainstorming is carried out in the form of a business game.

Scenario type methods. Methods for preparing and coordinating ideas about a problem or an analyzed object, set out in writing, are called scenarios. Initially, this method involved the preparation of a text containing a logical sequence of events or possible options solutions to problems over time. However, the obligatory requirement of time coordinates was later removed, and any document containing an analysis of the problem under consideration and proposals for its solution or for the development of the system, regardless of the form in which it is presented, began to be called a scenario. As a rule, in practice, proposals for the preparation of such documents are written by experts individually at first, and then an agreed text is formed.

The role of systems analysts in scenario preparation is to help recruited key experts in relevant areas of expertise to identify general patterns systems; analyze external and internal factors influencing its development and formation of goals; identify the sources of these factors; analyze the statements of leading experts in the periodical press, scientific publications and other sources of scientific and technical information; create auxiliary information funds (better automated) that contribute to the solution of the corresponding problem.

The scenario allows you to create a preliminary idea of the problem (system) in situations where it is not possible to immediately display it with a formal model. But still, a script is a text with all the ensuing consequences (synonymy, homonymy, paradoxes) associated with the possibility of its ambiguous interpretation by different specialists. Therefore, such a text should be considered as the basis for developing a more formalized view of the future system or problem being solved.

Methods of expert assessments. The basis of these methods is various forms of expert survey followed by evaluation and selection of the most preferred option. The possibility of using expert assessments, the justification of their objectivity is based on the fact that an unknown characteristic of the phenomenon under study is interpreted as a random variable, the reflection of the distribution law of which is an individual assessment of the expert on the reliability and significance of an event.

It is assumed that the true value of the characteristic under study is within the range of estimates received from the group of experts and that the generalized collective opinion is reliable. The most controversial point in these methods is the establishment of weighting coefficients according to the assessments expressed by experts and the reduction of conflicting assessments to some average value.

Expert survey This is not a one-time procedure. This way of obtaining information about a complex problem, characterized by a high degree of uncertainty, should become a kind of "mechanism" in a complex system, i.e. it is necessary to create a regular system of work with experts.

One of the varieties of the expert method is the method of studying the strengths and weaknesses of the organization, the opportunities and threats to its activities - the method of SWOT analysis.

This group of methods is widely used in socio-economic research.

Delphi type methods. Initially, the Delphi method was proposed as one of the brainstorming procedures and should help reduce the influence of psychological factors and increase the objectivity of expert assessments. Then the method began to be used independently. It is based on feedback, familiarizing the experts with the results of the previous round and taking these results into account when assessing the significance of the experts.

In specific methods that implement the "Delphi" procedure, this tool is used to varying degrees. So, in a simplified form, a sequence of iterative brainstorming cycles is organized. In a more complex version, a program of sequential individual surveys is developed using questionnaires that exclude contacts between experts, but provide for their acquaintance with each other's opinions between rounds. Questionnaires from tour to tour can be updated. To reduce factors such as suggestion or accommodation to the opinion of the majority, sometimes it is required that experts substantiate their point of view, but this does not always lead to the desired result, but, on the contrary, may increase the effect of adjustment. In the most advanced methods, experts are assigned weight coefficients of the significance of their opinions, calculated on the basis of previous surveys, refined from round to round, and taken into account when obtaining generalized assessment results.

Methods of the "tree of goals" type. The term "tree" implies the use of a hierarchical structure obtained by dividing the general goal into subgoals, and these, in turn, into more detailed components, which can be called subgoals of lower levels or, starting from a certain level, functions.

The "tree of goals" method is focused on obtaining a relatively stable structure of the goals of problems, directions, i.e. a structure that has changed little over a period of time with the inevitable changes that occur in any developing system.

To achieve this, when constructing the initial version of the structure, one should take into account the patterns of goal formation and use the principles of forming hierarchical structures.

Morphological methods. The main idea of the morphological approach is to systematically find all possible solutions to the problem by combining the selected elements or their features. In a systematic form, the method of morphological analysis was first proposed by the Swiss astronomer F. Zwicky and is often called the "Zwicky method".

business games - the simulation method has been developed for making managerial decisions in various situations by playing a group of people or a person and a computer according to the given rules. Business games allow, with the help of modeling and imitation of processes, to analyze, solve complex practical problems, ensure the formation of a thinking culture, management, communication skills, decision-making, instrumental expansion of managerial skills.

Business games act as a means of analyzing management systems and training specialists.

To describe management systems in practice, a number of formalized methods are used, which to varying degrees provide the study of the functioning of systems in time, the study of management schemes, the composition of units, their subordination, etc., in order to create normal working conditions for the management apparatus, personalization and clear information management

One of the most complete classifications, based on a formalized representation of systems, i.e. on a mathematical basis, includes the following methods:

- analytical (methods of both classical mathematics and mathematical programming);
- statistical ( mathematical statistics, probability theory, queuing theory);
- set-theoretic, logical, linguistic, semiotic (considered as sections of discrete mathematics);

graphic (graph theory, etc.).

The class of poorly organized systems corresponds in this classification to statistical representations. For the class of self-organizing systems, the most suitable models are discrete mathematics and graphical models, as well as their combinations.

Applied classifications are focused on economic and mathematical methods and models and are mainly determined by the functional set of tasks solved by the system.

Consider examples of system analysis:

Example . Consider a simple task- go to the university in the morning. This problem, often solved by a student, has all aspects:

- material, physical aspect - the student needs to move a certain mass, for example, textbooks and notebooks to the required distance;
- energy aspect - the student needs to have and spend a specific amount of energy to move;
- information aspect - information is needed about the route of movement and the location of the university, and it needs to be processed along the way of one's movement;
- human aspect - movement, in particular, movement by bus is impossible without a person, for example, without a bus driver;
- organizational aspect - suitable transport networks and routes, stops, etc. are needed;
- spatial aspect - moving a certain distance;
- time aspect - on given displacement time will be spent (during which there will be corresponding irreversible changes in the environment, in relations, in connections).

All types of resources are closely related and intertwined. Moreover, they are impossible without each other, the actualization of one of them leads to the actualization of the other.

Types of thinking

A special type of thinking is systemic, inherent in an analyst who wants not only to understand the essence of the process, phenomenon, but also to control it. Sometimes it is identified with analytical thinking, but this identification is not complete. An analytical mindset can be, and a systems approach is a methodology based on systems theory.

Subject (subject-oriented) thinking is a method (principle) with the help of which it is possible to purposefully (usually for the purpose of studying) identify and update, learn cause-and-effect relationships and patterns in a number of private and general events and phenomena. Often this is a technique and technology for studying systems.

Systemic (system-oriented) thinking is a method (principle) with the help of which it is possible to purposefully (usually for the purpose of management) identify and update, learn cause-and-effect relationships and patterns in a number of general and universal events and phenomena. It is often a systems research methodology.

In systems thinking, a set of events, phenomena (which may consist of various constituent elements) is updated, studied as a whole, as one event organized according to general rules, a phenomenon whose behavior can be predicted, predicted (as a rule) without clarifying not only the behavior of the constituent elements, but also the quality and quantity of themselves. Until it is understood how the system as a whole functions or develops, no knowledge of its parts will give a complete picture of this development.

Methodology is a system of principles and methods for organizing and building theoretical and practical activities. If the theory is the result of the process of cognition, then the methodology is the rationale for the way to achieve and build the knowledge obtained on its basis. The methodology provides a philosophical justification for the methods and techniques of organizing the entire variety of types (including cognitive) of human activity and involves the development of methods that are adequate to the objects being studied and transformed. One of the most important functions of methodology is heuristic: it should not only describe and explain a certain subject area, but at the same time be a tool for searching for new knowledge.

To put it briefly, methodology is the doctrine of method.

For social sciences can be identified three levels of methodology:

general scientific (for example, a systematic approach);
general social (social philosophy);
private social (sociology of personality, labor, youth

Method - a set of techniques and operations of theoretical and practical development of reality. For the field of social research, this is the main way to collect, process and analyze empirical materials.

Methodology - a set of technical techniques due to this method, including private operations, their sequence and interconnection.

IN modern science and social practice as a general scientific methodology, designed to formulate in a complete form a fairly universal set of research methods, as well as techniques and rules of constructive activity for subject areas of very different types and classes, is systems approach. The systems approach is based on principle of consistency, according to which the complex phenomena of objective reality are considered as integral phenomena formed by special mechanisms of communication and functioning of their constituent parts. On this basis, a specialized cognitive apparatus is formed, which determines the way of seeing the real world.

As you know, a system is such a set of interrelated elements, the interaction of which generates a special system quality, quite clearly localizing this set in the space surrounding it. It should be noted that the elements forming the system are attached to the specified system quality only as part of this system.

The system is always in a state of interaction with the external environment, which for it is, on the one hand, a source of resources necessary for its life, on the other hand, a source of various kinds of disturbing influences that can be useful (and then they are assimilated by the system), neutral (the system of their simply ignores) or harmful (the system tries to dampen their negative impact with the help and within the available resources).

Systematic consideration of an object involves:

definition and study of systemic quality;
identification of the totality of elements forming the system;
establishing links between these elements;
study of the properties of the environment surrounding the system, important for

the functioning of the system, at the macro and micro levels;

Revealing the relationships connecting the system with the environment.

The development of science and management practice also shows that a systematic approach to the study of a complex society makes it possible to comprehensively study the structural units of society (classes, layers, groups, associations, personalities), social relations between them (contacts, actions, interactions, social relations, social institutions), as well as the dynamics of social structures (social changes, processes).

The main advantage of the systematic approach is that it requires the maximum possible consideration of all aspects of the problem in their relationship and integrity, highlighting the main and essential, determining the nature and direction of the links between the structural components of the problem.

System analysis in a narrow sense, it is a set of scientific methods and practical techniques that can be used in the research and / or development of complex and super-complex objects, as well as in solving various problems that arise in all areas of managing social and organizational and technological systems. In a broad sense, systems analysis is understood as a synonym for a systems approach.

Scientific apparatus and methodological arsenal of system analysis in in general terms formed in the United States in the early 1940s. 20th century in the search for new approaches to solving the very complicated problems of production and rapid improvement of new types of weapons. It was noted that the main issue in solving any problems - regardless of their area, content and nature - is the choice of the most optimal solution alternative. However, this choice depends on the ability to evaluate the effectiveness of each alternative and the costs required for its implementation. Such operations were mastered by the investment of capital and the development of industry even before the Second World War. For their implementation, a number of methods were proposed, which, despite the constructiveness of the results in these areas, were almost never used in the field of armaments. Work on the creation of weapons systems began without considering how they would be used, how much they would cost, and whether their use would justify the costs of development and creation. The reason for this situation was that at that time the relative costs of armaments were low, there were few options for choice, so the principle of "nothing but the best" was actually used. During the Second World War, and especially with the beginning of the "atomic age", the cost of creating weapons increased many times over, and this approach became unacceptable. It was gradually replaced by another: "only what is needed, and at the minimum cost."

To implement this principle, it was necessary to be able to find, evaluate and compare simultaneously many alternatives for the production of weapons of various types. Operations research models developed by this time in industry and commerce could not be used for this because of their inherent limitations. New methods were required to be able to consider many alternatives, each of which was described a large number variables as a whole, while ensuring the completeness of the assessment of each alternative and the level of its uncertainty. The resulting universal problem-solving methodology was named by its authors "system analysis". The new methodology created to solve military problems was primarily used in this area. However, it soon became clear that civil, financial and many other problems of firms not only allow, but also require the use of this methodology.

System analysis quickly absorbed the achievements of many related and related fields and various approaches and turned into an independent scientific and applied discipline, rich in forms and areas of application, unique in its purpose and nature, and a field of professional activity.

The initial theoretical basis for system analysis is the theory of systems and the system approach. However, systems analysis borrows from them only the most general concepts and premises. In contrast, for example, to the systems approach, system analysis has a developed methodological and instrumental apparatus of its own and borrowed from other areas of science.

System analysis is based on strict adherence to the following principles:

the decision-making process should begin with a justification and a clear formulation of the final goals;
any problem should be presented as a whole one system indicating the relationships and consequences of each particular decision;
the solution of the problem should be represented by a set of possible alternative ways to achieve the goal;
the goals of individual units should not contradict the goals of the entire system as a whole.

The specificity of system analysis is an orientation towards finding optimal solutions with limited resources (personnel, finance, time, technology, etc.). It begins at the stage of the management cycle, when the goals of management are determined and ordered while finding a correspondence between the goals, possible ways to achieve them, the necessary and available resources for this.

In the center system analysis methodologies the operation of a quantitative comparison of alternatives is found, which is performed in order to select the optimal (according to certain criteria) alternative, which is supposed to be implemented. This can be achieved if all elements of the alternative are taken into account and the correct estimates are given to each of them. Thus, the idea arises of highlighting all the elements associated with a given alternative, i.e., "comprehensive consideration of all circumstances." The resulting integrity is called in system analysis complete system or simply system. The only criterion that makes it possible to single out this system can only be the fact of participation of this element in the process leading to the appearance of a given (target, desired) output result for a given alternative. Thus the concept process turns out to be central in the methodology of system analysis. There can be no systems thinking without a clear understanding of the process.

To define a system means to define system objects, their properties and relationships. The most important of these are input, process, output, feedback, and constraint.

System input what is changed during a given process is called. Or otherwise, this is what this process must be applied to in order to obtain the desired result. In many cases, the components of an input are a "working input" (what is being "processed") and a processor (what is being "processed"). System output called the result or final state of the process. The process translates an input into an output. The ability to transform an input into a specific output is called property of a given process or transfer function (IV).

Here it is necessary to pay attention to the fact that social world processes do not always translate "input" into definite"exit" due to the fact that social structures are not at all similar to those "devices" that are considered in classical system models. Unlike the latter, which work out input signals on rigid (or non-rigid, but quite predictable, probabilistic) algorithms social structures, being predominantly self-organizing systems, only perceive management influences. But far from passive and highly subjective. For this reason, they cannot be displayed in formal constructions using fixed transfer functions that indicate the nature of the transformation of "input" into "output". Social objects are constantly changing, perceiving and associating in the most bizarre way all the phenomena of internal and external order that are somehow significant.

In any functioning system, there are three sub-processes of different roles: the main process, feedback and restriction. The main process converts input to output. Feedback performs a number of operations: compares the real state of the output with a given (target) model and highlights the difference (A). The subsequent analysis of the content and meaning of the difference makes it possible to develop, if necessary, a managerial decision. The need for a decision arises when the difference in the state of input and output exceeds some set or accepted level, that is, when a problem arises for which a decision must be made. The meaning of this solution lies in such a correction of the system process, the implementation of which could bring the real state of the system output closer to its model or bring their difference to an acceptable level.

Limitation there is a sum of rules, regulations, and guidelines put forward personally or from outside, defining the boundary of the problem. It is formed by the consumer (buyer) of the system output. In a generalized form, the constraint can be viewed as external environment generally. The system constraint is taken into account when making a management decision, ensuring that the system output matches the consumer's goals. Thus, the constraint of the system is reflected in the adjusted output model.

The functioning system is shown in fig. 2.1. A circle with an oblique cross denotes a comparison unit (comparator, adder), in which all the most important controlled parameters are compared.

Rice. 2.1.

In systems analysis, it is postulated that every system consists of subsystems and every system is a subsystem of some other system of a higher order. It is also postulated that any system can be described in terms of system objects, properties and relationships. The system boundary is determined by a set of inputs from the external environment. The external environment is a set of systems for which this system is not a functional subsystem.

problem called a situation characterized by a difference between the necessary (desired) and the existing outputs. The latter is necessary if its absence creates a threat to the existence or development of the system. It is provided by the existing system. The desired output is provided by the desired system. The problem is the difference between the existing and desired systems. The problem may be to prevent a decrease or to increase the output. The problem condition represents the existing system (the "known"). The requirement represents the desired system. Solution to the problem there is something that fills the gap between the existing and the desired systems. The system that fills the gap is the object of construction.

Problems can manifest themselves in symptoms. Systematically manifested symptoms form trend. Finding a problem is the result of a process of identifying symptoms. Identification is possible under the condition of knowledge of the norm or the desired behavior of the system. The detection of a problem is followed by the prediction of its development and the assessment of the relevance of its solution, i.e., the state of the system with an unresolved problem. Assessment of the relevance of solving the problem allows you to determine the need for its solution.

The process of finding a solution centers around the iteratively performed operations of identifying the condition, goal, and possibilities for solving the problem. The result of identification is a description of the condition, purpose and capabilities in terms of system objects (input, process, output, feedback and limitation), properties and relationships. If the structures and elements of the conditions, goals and possibilities of this problem are known, identification has the character of determining quantitative relationships, and the problem is called quantitative. If the structure and elements of the conditions, goals and opportunities are known in part, the identification is qualitative, and the problem is called qualitative or semi-structured. As a problem-solving methodology, systems analysis indicates a fundamentally necessary sequence of interrelated operations, which (in the most general terms) consists of identifying a problem, constructing a solution to the problem, and implementing this solution. The decision process is the design, evaluation and selection of system alternatives according to the criteria of cost, time, efficiency and risk, taking into account the relationship between the marginal increments of these quantities (marginal ratios). The choice of the boundaries of this process is determined by the condition, purpose and possibilities of its implementation. The most adequate construction of this process involves the comprehensive use of heuristic conclusions within the framework of the postulated structure of the system methodology.

The reduction of the number of variables is carried out on the basis of an analysis of the sensitivity of the problem to changes in individual variables or groups of variables, aggregation of variables into summary factors by selecting criteria of an appropriate form, and also using, where possible, mathematical methods for reducing enumeration (mathematical programming, etc.). The logical integrity of the process is provided by explicit or implicit assumptions, each of which can be a source of risk. It is postulated that the structure of system functions and problem solving is standard for any systems and any problems. Only the methods of implementing functions can change. Improving methods for this condition scientific knowledge has a limit, defined as a potentially achievable level. As a result of solving the problem, new connections and relationships are established, some of which determine the desired outcome, and the other determines unforeseen opportunities and limitations that can become a source of future problems.

INTRODUCTION

System analysis is a scientific discipline that deals with solving problems related to the study of systems of various physical nature, purpose and scale, managing the evolution of systems, optimizing parameters, structure and algorithms for the functioning of systems, making optimal decisions on the organization and development of systems. Therefore, the origins of systems analysis and its methodology lie in systems theory, operations research theory, decision theory, and control theory.

The emergence of the discipline "system analysis" is due to the need to conduct research on systems of an interdisciplinary nature. Creation, operation and development of complex technical systems, design and management of large-scale energy, transport, production systems, analysis ecological systems and systems of social purpose and many other areas of practical and scientific activity required the organization of research that would be of an unconventional nature.

On the present stage development of system analysis, its apparatus and tools are based on the widespread use of computers and include a complex and developed system of models. The development of system analysis was determined, on the one hand, by the development of the mathematical apparatus and the development of formalization methods, and on the other hand, by new tasks that arise in industry, economics, military affairs, etc. System analysis includes both Scientific research systems, as well as relevant activities aimed at the practical implementation of the results of such studies.

The scientific discipline called systems analysis studies events and processes in systems, develops models designed to explain these events and processes, and uses these models to study changes in the evolution and characteristics of systems when its structural and functional parameters change. Thus, systems analysis is a science, since this discipline uses scientific method to obtain relevant knowledge and differs from other sciences by the subject of research. System analysis, like any other science, requires the development of its own mathematical apparatus of methods of system analysis, focused on the specifics inherent in this area and research objectives.

Distinctive features of system analysis are that it is based on the use of a modern scientific approach to the study and management of systems of various nature and purpose - the system principle, integrated research teams and the scientific method.

for solving problems of system analysis. The systems principle is the recognition that every system is made up of parts, each with its own evolutionary goals, and that in any system the evolution of each part affects all other parts of the system. The scientific method of system analysis, in particular, is based on the fact that, as a rule, the entire system that is the object of study cannot be subjected to a natural experiment. Therefore, in most cases, investigating the system

in In general, it is necessary to apply an approach that is not related to conducting full-scale experiments.

The concept of the systems principle has had a significant impact on the planning and executive functions of systems management. System administrators are increasingly turning to systems analysts for help in choosing from a variety of possible solutions. The value of the system principle for managing the system is determined by the content of the main goal of management. First, it is necessary to achieve the efficiency of the functioning of the system as a whole and not allow the interests of any one part of the system to interfere with the achievement of the overall goals of the creation and functioning of the system. Secondly, it is necessary to achieve this on the condition that the parts of the system have, as a rule, conflicting goals for their functioning. Thirdly, it is necessary to understand that it is possible to achieve the general goals of the functioning of the system only if it is considered as a whole, striving for this to understand and evaluate the interaction of all its parts and combine them on such a basis that would allow the system as a whole to effectively achieve her goal. Any formal analysis of the system, or even an attempt at a formal analysis, is usually valuable in that, at a minimum, it makes the system administrator think about the main thing and move

in direction. And although the system analyst in his conclusion will not always be able to accurately indicate to the administrator which solution would be the best, the very fact of the analysis will require him to list the alternatives and formulate the goals of the system analysis.

Without striving for an exhaustive formal definition of system analysis, we note that this science is mainly engaged in the analysis of organizational (functional) systems, i.e. systems whose work is determined by the decisions of people (as opposed to, for example, physical systems that obey only the laws of nature) . System analysis provides a mathematical description of the processes of functioning of systems and their management. It is focused on solving problems for which you can build mathematical models systems for optimal solutions. In any system analysis project, the following main stages can be distinguished: problem statement, development of a system model, finding a solution, checking the model and evaluating the solution, implementing the solution and monitoring its correctness. In sys-

dark analysis the main role assigned mathematical modeling. To build a mathematical model, it is necessary to have a clear understanding of the purpose of the functioning of the system under study and to have information about the restrictions that determine the range of acceptable values of the controlled variables. Analysis of the model should lead to the determination of the best impact on the object of study if all the established restrictions are met.

The complexity of real systems can make it very difficult to present the goal and constraints in an analytical way. Therefore, it is very important to reduce the "dimension" of the problem being solved in such a way as to ensure the possibility of constructing an appropriate model. Despite too big number variables and constraints that, at first glance, must be taken into account when analyzing real systems, only a small part of them turns out to be essential for describing the behavior of the systems under study. Therefore, in a simplified description of real systems, on the basis of which one or another model will be built, one should first of all identify the essential variables, parameters, and limitations.

When the term "systems analysis" is used, it almost always means the application of mathematical methods to model systems and analyze their characteristics. Indeed, mathematical models and methods occupy a central place in system analysis. However, it should be borne in mind that solving problems of organizational management does not always come down to building models and performing appropriate experiments with them. This is due, in particular, to the fact that in the course of the formation of control decisions one often encounters factors that are essential for the correct solution of the problem, but are not amenable to strict formalization and, therefore, cannot be directly introduced into the mathematical model. One of the hard-to-formalizable factors of this kind is the factor of human activity.

System analysis as a methodology for solving problems of research and management of systems can be considered both as a science and as an art. The scientific content of system analysis is provided by the effective use of mathematical models and methods in solving problems of research and control of systems. At the same time, the successful completion of all stages of the study, from its beginning to the implementation of the solution obtained using the developed mathematical model, is largely determined by creativity and intuition of researchers.

PROBLEMS OF SYSTEM ANALYSIS

1.1. Systems and Models

A system is a set of objects together with relationships between objects and between their attributes.

This definition assumes that a system has properties, functions, and purposes that are distinct from those of its constituent objects, relationships, and attributes.

Objects are simply parts or components of a system. Most of the systems that surround or interest us are

from physical parts, however, abstract objects can also be included in systems: mathematical variables, equations, laws, etc.

Attributes are properties of objects.

Attitude is one of the forms of the universal interconnection of all objects, phenomena, processes in nature, society and thinking.

The relations of objects to each other are extremely diverse: cause and effect, part and whole, the relationship between parts within the whole, argument and function, etc. In mathematics and logic, such types of relations as “... more than ... ”, “... implies ...”, etc. Any set of objects has internal relations, because the distance between objects can always be taken as a relation. It is assumed that the relations considered in a certain context depend on the problem being solved, and on this basis, certain essential or interesting relations are included in the consideration and trivial or non-essential relations are excluded. Researcher, problem-solving, he decides which relations are essential and which are trivial.

System environment- a set of all objects whose attributes or relationships change affect the system, as well as those objects whose attributes or relationships between these objects change as a result of the system.

The above definition raises a natural question: when is an object considered to belong to the environment, and when does it belong to the system? If some object interacts with the system in the way specified in the definition, does this mean that it is part of the system? The answers to these questions are not obvious. In the famous

sense, the system, together with the environment, represents a set of objects that are of interest to the researcher in specific task. The division of this set into two sets - the system and the environment - can be done different ways, and all of them are quite arbitrary. Ultimately, the solution to this problem depends on the goals of the one who considers a certain set of objects as a system.

General definition problem environment this system is far from simple. In order to fully define the environment, one must know all the factors that affect the system or are determined by the system. As a rule, the researcher includes in the composition of the system and its environment all those objects that seem to him the most important, describes the internal relations of the system as completely as possible, and pays more attention to its most important properties, neglecting those properties that, in his opinion, opinion do not play a significant role. This idealization method is widely used, for example, in physics and chemistry. Biologists, sociologists, economists, and other scientists interested in living systems and their behavior are in a more difficult position. In these sciences it is very difficult to distinguish the essential variables of systems from the non-essential ones; in other words, the problem of specification of the studied set of objects and its subsequent division into two sets - the system and the environment - is here a fundamental difficulty.

From the definition of system and environment, it follows that any system can be divided into subsystems. Objects belonging to one subsystem can be considered as parts of the environment of another subsystem. The analysis of a subsystem requires, of course, the consideration of a new set of relations. Of course, the behavior of a subsystem cannot be completely analogous to the behavior of the system that includes it. In particular, such a property of systems as the hierarchical ordering of the system, in fact, reflects the possibility of dividing the system into subsystems. In other words, it can be said that parts of a system can themselves be systems of lower orders. One method of studying a complex system is to examine in detail the behavior of one of its subsystems. Another method is to observe only the macroscopic behavior of the system as a whole. Both of these methods are widely used in various fields of knowledge, and both of them are important.

In the definition of the system, it is noted that all systems are characterized by the presence of relationships between objects and between their attributes.

If every part of the system is so related to every other part that a change in some part causes a change in all other parts.

tyakh and in the whole system as a whole, then the system behaves as an integrity, or as some connected formation.

If in a set of completely unrelated objects, a change in each part of the set depends only on this part itself, and the change in the set as a whole is the physical sum of changes in its individual parts, then such a set is called isolated or physically additive.

Integrity and isolation, obviously, are not two different properties, but the limiting values of some measure of the same property. Integrity and separateness differ in the degree to which this property is present, and there is currently no method to measure them. The term "complex" is often used to describe a set of parts that are independent of each other, and the term "system" is used only when a certain degree of integrity is characteristic of a set of objects. However, it is more correct to use the term "degenerate system" for a set of completely independent parts.

Modeling is the replacement of one system (original) with another (model) and the study of the properties of the original by examining the properties of the model. Substitution is made in order to simplify the study of the properties of the original.

In general, the original system can be any natural or artificial, real or abstract system. It has a certain set of parameters and is characterized by certain properties. The system manifests its properties under the influence of external influences. The set of system parameters and their values reflects its internal content - composition, structure and functioning algorithms. The set and values of parameters distinguish the system from other systems. The characteristics of the system are mainly its external features, which are important when interacting with other systems. The characteristics of the system are functionally dependent on its parameters. Obviously, each characteristic of the system is determined mainly by a limited subset of parameters. It is assumed that the influence of other system parameters on the value of this characteristic of the system can be neglected. As a rule, researchers are only interested in certain characteristics of the system under study under specific external influences on the system.

A model is also a system with its own sets of parameters and characteristics, correspondingly reflecting the sets of parameters and characteristics of the original system. With some approximation, we can assume that the characteristics of the model are related to the characteristics of the original.

In this case, the set of characteristics of the model is a reflection of the set of interesting characteristics of the original. Modeling is advisable when the model does not have those features of the original that prevent its study, or there are parameters different from the original that contribute to the study of the properties of the model.

Modeling theory is an interconnected set of provisions, definitions, methods and tools for creating and studying models. These provisions, definitions, methods and means, as well as the models themselves, are the subject of modeling theory. The main task of modeling theory is to equip researchers with a methodology for creating such models that accurately and completely capture the properties of interest of the originals, are easier or faster to study and ensure the use of its results to obtain the necessary data on the characteristics of the simulated system originals. Modeling theory is the main component of the general theory of systems - systemology, in which the feasibility of models is postulated as the main principle: the system is represented by a finite set of models, each of which reflects a certain facet of its essence.

1.2. System classification

When considering systems, you can use various ways to classify them: by origin, according to the description of input and output

variables, according to the description of the system operator, according to the type of control.

On fig. 1.1 shows a diagram of a two-level classification of systems by origin. If the completeness of the classification of the first level is logically clear, then the second level is clearly incomplete. The classification of natural systems is clear from the figure, its incompleteness is obvious. The incompleteness of the division of artificial systems is associated, for example, with the still unfinished development of artificial intelligence systems. Examples of subclasses of mixed systems include ergonomic systems (machine-human-operator complexes), biotechnical systems (systems that include living organisms and technical devices), and organizational systems (consisting of teams of people who are equipped with the necessary technical means).

S Y S T E M S

NATURAL

ARTIFICIAL

MIXED

Mechanisms

Ergonomic

Biotechnical

Environmental

Automata

Organizational

Social

. . . . . . . . . . . . . . .

. . . . . . . . . . . .

Rice. 1.1. Classification of systems by origin.

The three-level classification scheme of systems according to the type of input, output and internal variables is shown in fig. 1.2. There is a fundamental difference between variables described qualitatively and quantitatively, which is the basis of the first level of classification. For completeness, a third class has been introduced; it includes systems in which some of the variables are of a qualitative nature, and the rest are quantitative. At the next level of classification of systems with qualitative variables, there are cases where the description is carried out by means of a natural language, and cases that allow deeper formalization. The second level of classification of systems with quantitative variables is caused by differences in the methods of discrete and continuous mathematics, which is reflected in the names of the introduced subclasses; the case is also envisaged when the system has both continuous and discrete variables. For systems with a mixed quantitative-qualitative description of variables, the second level is the union of subclasses of the first two classes and is not shown in the figure. The third level of classification is the same for all subclasses of the second level and is depicted for only one of them.

S Y S T E M S

WITH QUALITY

WITH QUANTITATIVE

WITH MIXED

VARIABLES

DESCRIPTION

VARIABLES

description

Discrete

formalized

description

continuous

mixed

description

mixed

deterministic

Stochastic

mixed

Rice. 1.2. A fragment of the classification of systems according to the description of variables.

The next classification (Fig. 1.3) is by the type of the system operator, i.e., the classification of the types of relationships between input and output variables.


		S Y S T E M S

	NONPARAMETER-		PARAMETERS-	WHITE BOX
	DRAWED		CALLED	(operator
(operator				known
unknown)	(operator		(operator	fully)
	known		known
	partially)		to parameters)

Inertial (with memory)

Inertialess (no memory)

Closed (with feedback)

Open (without feedback)

Linear

Nonlinear

Quasilinear

Rice. 1.3. A fragment of the classification of systems according to the type of operators.

At the first level, there are classes of systems that differ in the degree of availability of information about the system operator. The branch of the "black box" ends at this level: the operator is generally considered unknown. The more information about the operator is available, the more differences can be considered and the more developed the classification will be. For example, information about the operator can be so general character that the description of the system cannot be obtained in a parametrized functional form. A non-parameterized class of systems and fits similar situations with very limited information about the operator.

Our knowledge about the operator can have a level that allows us to make a parametric description of this operator, i.e., write down the dependence of the system output y (t) on the system input x (t) in explicit form up to a finite number of parameters θ = (θ 1 , K , θ k ) : y (t ) = Φ (x (), θ ) , where Φ denotes the system operator. Such systems belong to the third class in the classification of this species.

Finally, if the operator parameters are specified exactly, then any uncertainty disappears and we have a system with a fully defined operator, i.e., a “white box”.

Further levels of classification in fig. 1.3 are given only for systems of the third and fourth classes (“black box” is not subject to

further classification, and the classification of non-parametrized systems is related to the type of information available about their operators). The second, third and fourth levels are clear from the drawing itself. Of course, the classification can be continued (for example, linear operators are usually divided into differential, integral, etc.).

Considering the output y (t) of the system (it can be a vector) as its response to controlled u (t) and uncontrolled w (t) inputs - x (t) = (u (t), w (t)), the “black box" can be represented as a set of two processes: X = (x (t ), t T ) and Y = ( y (t ), t T ) . If we consider y (t ) the result of some transformation Φ of the process x (t ) , i.e. y (t ) = Φ (x (t )) , then the "black box" model assumes that this transformation is unknown. In the same case, when we are dealing with a "white box", the correspondence between input and output can be described in one way or another. Which way depends on what we know and in what form this knowledge can be used.

The scheme of the next method of classifying systems - by type of control - is shown in fig. 1.4. The first level of classification is determined by whether the control unit is included in the system or is external to it; a class of systems is also distinguished, the control of which is divided and partially carried out from the outside, and partially - within the system itself. Regardless of whether the control block is included in the system or removed from it, four main types of control can be distinguished, which is reflected at the second level of classification. These types differ depending on the degree of availability of information about the trajectory of the system in the state space, leading the system to the goal, and the ability of the control unit to ensure the evolution of the system along this trajectory.

S Y S T E M S

WITH EXTERNAL

SELF-GOVERNED

WITH COMBINED

MANAGEMENT

no feedback

Program control

Automatic

Regulation

Automatic control

semi-automatic

Control

Parametric adaptation

automated

by parameters

Control

Structural adaptation

Organizational

by structure

(self-organization)

Rice. 1.4. Classification of systems by type of control.

System analysis- a scientific method of cognition, which is a sequence of actions to establish structural relationships between variables or elements of the system under study. It is based on a set of general scientific, experimental, natural science, statistical, and mathematical methods.

To solve well-structured quantifiable problems, the well-known methodology of operations research is used, which consists in constructing an adequate mathematical model (for example, linear, nonlinear, dynamic programming problems, problems of queuing theory, game theory, etc.) and applying methods to find the optimal control strategy targeted actions.

System analysis provides the following system methods and procedures for use in various sciences, systems:

abstraction and specification

analysis and synthesis, induction and deduction

Formalization and concretization

composition and decomposition

Linearization and selection of non-linear components

Structuring and restructuring

· prototyping

reengineering

algorithmization

simulation and experiment

software control and regulation

Recognition and identification

clustering and classification

expert evaluation and testing

verification

and other methods and procedures.

It should be noted the tasks of studying the system of interactions of the analyzed objects with the environment. The solution to this problem involves:

- drawing a boundary between the system under study and the environment, which determines the maximum depth

the influence of the interactions under consideration, to which the consideration is limited;

- determination of the real resources of such interaction;

– consideration of the interactions of the system under study with a higher level system.

Tasks of the following type are associated with the design of alternatives for this interaction, alternatives for the development of the system in time and space. An important direction in the development of systems analysis methods is associated with attempts to create new possibilities for constructing original solution alternatives, unexpected strategies, unusual ideas and hidden structures. In other words, speech here about the development of methods and means amplification of inductive capabilities human thinking in contrast to its deductive capabilities, which, in fact, are aimed at strengthening the development of formal logical means. Research in this direction has begun only quite recently, and there is still no single conceptual apparatus in them. Nevertheless, several important areas can be distinguished here, such as the development the formal apparatus of inductive logic, methods of morphological analysis and other structural and syntactic methods for constructing new alternatives, syntactic methods and organization of group interaction in solving creative problems, as well as the study of the main paradigms of search thinking.

Tasks of the third type consist in constructing a set simulation models describing the influence of one or another interaction on the behavior of the object of study. It should be noted that system studies do not pursue the goal of creating some kind of supermodel. We are talking about the development of private models, each of which solves its own specific issues.

Even after such simulation models created and investigated, the question of bringing together various aspects of the system's behavior into a single scheme remains open. However, it can and should be solved not by building a supermodel, but by analyzing the reactions to the observed behavior of other interacting objects, i.e. by studying the behavior of objects - analogues and transferring the results of these studies to the object of system analysis. Such a study provides a basis for a meaningful understanding of situations of interaction and the structure of relationships that determine the place of the system under study in the structure of the supersystem, of which it is a component.

Tasks of the fourth type are associated with the design decision making models. Any system study is connected with the study of various alternatives for the development of the system. The task of system analysts is to choose and justify the best development alternative. At the stage of development and decision-making, it is necessary to take into account the interaction of the system with its subsystems, combine the goals of the system with the goals of the subsystems, and single out global and secondary goals.

The most developed and at the same time the most specific area of scientific creativity is associated with the development of the theory of decision making and the formation of target structures, programs and plans. There is no lack of work and actively working researchers here. However, in this case, too many results are at the level of unconfirmed inventions and discrepancies in understanding both the essence of the tasks and the means to solve them. Research in this area includes:

a) building a theory for evaluating the effectiveness of decisions made or plans and programs formed;

b) solving the problem of multi-criteria in the evaluation of decision or planning alternatives;

c) study of the problem of uncertainty, especially associated not with statistical factors, but with the uncertainty of expert judgments and deliberately created uncertainty associated with simplifying ideas about the behavior of the system;

d) development of the problem of aggregating individual preferences on decisions affecting the interests of several parties that affect the behavior of the system;

e) study of specific features of socio-economic criteria of efficiency;

f) creation of methods for checking the logical consistency of target structures and plans and establishing the necessary balance between the predetermination of the action program and its readiness for restructuring when a new one arrives

information about both external events and changes in ideas about the execution of this program.

The latter direction requires a new awareness of the real functions of the target structures, plans, programs and the definition of those that they should perform, as well as the links between them.

The considered tasks of system analysis do not cover the full list of tasks. Listed here are those that present the greatest difficulty in solving them. It should be noted that all the tasks of systemic research are closely interconnected with each other, cannot be isolated and solved separately, both in time and in terms of the composition of performers. Moreover, in order to solve all these problems, the researcher must have a broad outlook and possess a rich arsenal of methods and means of scientific research.

ANALYTICAL AND STATISTICAL METHODS. These groups of methods are most widely used in the practice of design and management. True, graphical representations (graphs, diagrams, etc.) are widely used to present intermediate and final results of modeling. However, the latter are auxiliary; the basis of the model, the proofs of its adequacy, are those or other directions of analytical and statistical representations. Therefore, despite the fact that in the main areas of these two classes of methods, universities read independent courses lectures, we nevertheless briefly characterize their features, advantages and disadvantages from the point of view of the possibility of using them in system modeling.

Analytical in the classification under consideration, methods are named that display real objects and processes in the form of points (dimensionless in strict mathematical proofs) that make any movements in space or interact with each other. The basis of the conceptual (terminological) apparatus of these representations is the concepts of classical mathematics (value, formula, function, equation, system of equations, logarithm, differential, integral, etc.).

Analytical representations have a long history of development, and they are characterized not only by the desire for rigor of terminology, but also by assigning certain letters to some special quantities (for example, doubling the ratio of the area of a circle to the area of a square inscribed in it p» 3.14; the base of the natural logarithm – e» 2.7, etc.).

On the basis of analytical ideas, there have arisen and are developing mathematical theories of varying complexity - from the apparatus of the classical mathematical analysis(methods for studying functions, their form, methods of representation, searching for extrema of functions, etc.) to such new sections of modern mathematics as mathematical programming (linear, nonlinear, dynamic, etc.), game theory (matrix games with pure strategies, differential games, etc.).

These theoretical directions have become the basis of many applied ones, including the theory of automatic control, the theory of optimal solutions, etc.

When modeling systems, a wide range of symbolic representations is used, using the "language" of classical mathematics. However, these symbolic representations do not always adequately reflect real complex processes, and in these cases, generally speaking, they cannot be considered rigorous mathematical models.

Most of the areas of mathematics do not contain the means of setting the problem and proving the adequacy of the model. The latter is proved by experiment, which, as the problems become more complex, also becomes more and more complex, expensive, not always indisputable and realizable.

At the same time, this class of methods includes a relatively new area of mathematics - mathematical programming, which contains the means of setting the problem and expands the possibilities of proving the adequacy of models.

Statistical representations formed as an independent scientific direction in the middle of the last century (although they arose much earlier). They are based on the display of phenomena and processes using random (stochastic) events and their behavior, which are described by the corresponding probabilistic (statistical) characteristics and statistical patterns. Statistical mappings of the system in the general case (by analogy with analytical ones) can be represented as if in the form of a “blurred” point (fuzzy area) in n-dimensional space, into which the system (its properties taken into account in the model) is transferred by the operator F. “Blurred” point should be understood as a certain area characterizing the movement of the system (its behavior); in this case, the boundaries of the region are given with a certain probability p (“blurred”) and the movement of the point is described by some random function.

Fixing all the parameters of this area, except for one, you can get a cut along the line a - b, the meaning of which is the impact of this parameter on the behavior of the system, which can be described by a statistical distribution for this parameter. Similarly, you can get two-dimensional, three-dimensional, etc. statistical distribution patterns. Statistical regularities can be represented as discrete random variables and their probabilities, or as continuous dependences of the distribution of events and processes.

For discrete events, the relation between possible values random variable xi and their probabilities pi are called the distribution law.

Brainstorming method

A group of researchers (experts) develops ways to solve the problem, while any method (any thought expressed aloud) is included in the number of considered ones, the more ideas, the better. At the preliminary stage, the quality of the proposed methods is not taken into account, that is, the subject of the search is the creation of as many options for solving the problem as possible. But to be successful, the following conditions must be met:

the presence of an inspirer of ideas;

· a group of experts does not exceed 5-6 people;

· the potential of researchers is commensurable;

the environment is calm;

equal rights are observed, any solution can be proposed, criticism of ideas is not allowed;

· Duration of work no more than 1 hour.

After the "flow of ideas" stops, the experts carry out a critical selection of proposals, taking into account the limitations of the organizational and economic nature. The selection of the best idea can be carried out according to several criteria.

This method is most productive at the stage of developing a solution for the implementation of the goal, when revealing the mechanism of the system's functioning, when choosing a criterion for solving the problem.

The method of "concentration of attention on the goals of the problem"

This method consists in selecting one of the objects (elements, concepts) associated with the problem being solved. At the same time, it is known that the object accepted for consideration is directly related to the ultimate goals of this problem. Then the connection between this object and some other, chosen at random, is examined. Next, the third element is selected, just as randomly, and its relationship with the first two is examined, and so on. Thus, a certain chain of interconnected objects, elements or concepts is created. If the chain breaks, then the process resumes, a second chain is created, and so on. This is how the system is explored.

Method "inputs-outputs of the system"

The system under study is necessarily considered together with the environment. In this case, special attention is paid to the restrictions that the external environment imposes on the system, as well as the restrictions inherent in the system itself.

At the first stage of studying the system, possible outputs of the system are considered and the results of its functioning are evaluated according to changes in the environment. Then the possible inputs of the system and their parameters are investigated, which allow the system to function within the limits of the accepted restrictions. And, finally, at the third stage, acceptable inputs are chosen that do not violate the system's limitations and do not bring it into conflict with the goals of the environment.

This method is most effective at the stages of understanding the mechanism of the system functioning and decision making.

Scenario method

The peculiarity of the method is that a group of highly qualified specialists in a descriptive form represents the possible course of events in a particular system - starting from the current situation and ending with some resulting situation. At the same time, artificially erected, but arising in real life restrictions on the entry and exit of the system (on raw materials, energy resources, finance, and so on).

The main idea of this method is to identify the links between various elements of the system that manifest themselves in a particular event or constraint. The result of such a study is a set of scenarios - possible directions for solving the problem, from which, by comparing according to some criterion, the most acceptable ones could be chosen.

Morphological method

This method involves the search for all possible solutions to the problem by exhaustive census of these solutions. For example, F.R. Matveev identifies six stages in the implementation of this method:

the formulation and definition of the constraints of the problem;

search for possible decision parameters and possible variations of these parameters;

Finding all possible combinations of these parameters in the resulting solutions;

Comparison of decisions in terms of the goals pursued;

Choice of solutions

· in-depth study of selected solutions.

Modeling methods

A model is a system created to represent a complex reality in a simplified and understandable form, in other words, a model is an imitation of this reality.

The problems solved by models are many and varied. The most important of them:

· with the help of models, researchers try to better understand the course of a complex process;

· with the help of models, experimentation is carried out in the case when this is not possible on a real object;

· with the help of models, the possibility of implementing various alternative solutions is evaluated.

In addition, models have such valuable properties as:

reproducibility by independent experimenters;

· variability and the possibility of improvement by introducing new data into the model or modifying relationships within the model.

Among the main types of models, symbolic and mathematical models should be noted.

Symbolic models - diagrams, diagrams, graphs, flowcharts and so on.

Mathematical models are abstract constructions that describe in mathematical form the connections, relationships between the elements of the system.

When building models, the following conditions must be observed:

have a sufficiently large amount of information about the behavior of the system;

Stylization of the functioning mechanisms of the system should take place within such limits that it would be possible to accurately reflect the number and nature of the relationships and connections existing in the system;

The use of automatic information processing methods, especially when the amount of data is large or the nature of the relationship between the elements of the system is very complex.

However, mathematical models have some disadvantages:

the desire to reflect the process under study in the form of conditions leads to a model that can be understood only by its developer;

On the other hand, simplification leads to a limitation of the number of factors included in the model; consequently, there is an inaccuracy in the reflection of reality;

· the author, having created a model, "forgets" that he does not take into account the action of numerous, maybe insignificant factors. But the combined effect of these factors on the system is such that the final results cannot be achieved on this model.

In order to level these shortcomings, the model must be checked:

How realistically and satisfactorily does it reflect the real process?

· whether changing the parameters causes a corresponding change in the results.

Complex systems, due to the presence of many discretely functioning subsystems, as a rule, cannot be adequately described using only mathematical models, so simulation modeling has become widespread. Simulation models have become widespread for two reasons: firstly, these models allow the use of all available information (graphic, verbal, mathematical models ...) and, secondly, because these models do not impose strict restrictions on the input data used. Thus, simulation models allow you to creatively use all the available information about the object of study.