Lecture
Everything that human activity is directed towards is called an object (Lat. objection — a thing). The development of methodology is aimed at
ordering the acquisition and processing of information about objects that exist outside our consciousness and interact with each other and with the external
environment.
In scientific research, hypotheses play a major role, i.e. certain predictions based on a small amount of experimental data, observations, and guesses. A quick and complete verification of proposed hypotheses can be carried out in the course of a specially designed experiment. In formulating and verifying the correctness of hypotheses, analogy is of great importance as a method of reasoning. An analogy is a judgement about some particular similarity of two objects, and such similarity may be either essential or inessential. It should be noted that the notions of the essentiality and inessentiality of the similarity or difference between objects are conventional and relative. The essentiality of a similarity (difference) depends on the level of abstraction and, in general, is determined by the ultimate purpose of the research being conducted. A modern scientific hypothesis is, as a rule, created by analogy with scientific propositions that have been verified in practice. Thus, analogy links the hypothesis with the experiment. Hypotheses and analogies that reflect the real, objectively existing world must be intuitive or reducible to logical schemes convenient for research; such logical schemes, which simplify reasoning and logical constructions or make it possible to carry out experiments clarifying the nature of phenomena, are called models.
In other words, a model (Lat. modulus — a measure) is a substitute object for the original object, enabling the study of some properties of the original.
Substituting one object for another in order to obtain information about the most important properties of the original object by means of the model object
is called modelling. Thus, modelling can be defined as the representation of an object by a model in order to obtain information about
that object by carrying out experiments with its model. The theory of substituting some objects (originals) with other objects (models) and studying the properties of objects using their models is called the theory of modelling.
If the results of modelling are confirmed and can serve as a basis for predicting the processes occurring in the objects under study, the model is said to be adequate to the object. In this case, the adequacy of the model depends on the purpose of modelling and the criteria adopted. In general terms, modelling can be defined as a method of cognition in which the object-original being studied is in a certain correspondence with another object-model, the model being capable, in one respect or another, of substituting for the original at certain stages of the cognitive process. The stages of cognition at which such substitution occurs, as well as
the forms of correspondence between the model and the original, can vary:
1) modelling as a cognitive process involving the processing of information arriving from the external environment about the phenomena taking place in it, as a result of which images corresponding to objects appear in the consciousness;
2) modelling that consists in constructing a certain system-model (a second system) connected by certain relations of similarity with
the original system (the first system), in which case the mapping of
one system onto another serves as a means of revealing the dependencies between
the two systems, reflected in the relations of similarity, rather than as a result of
the direct study of the incoming information.
When modelling an information system, it is necessary to take into account
the following features: the complexity of the structure and the stochastic nature of the connections between
elements, the ambiguity of behaviour algorithms under different conditions,
the large number of parameters and variables, the incompleteness and
indeterminacy of the initial information, the diversity and probabilistic
nature of the effects of the external environment, and so on. The limited possibilities of
experimental research on large systems make it important to
develop a methodology for their modelling that would make it possible, in a
suitable form, to represent the processes of system operation,
to describe the course of these processes by means of mathematical models,
and to obtain the results of experiments with the models for evaluating the characteristics
of the objects under study. Moreover, at different stages of the creation and use
of the listed systems, for the whole variety of subsystems included in them,
the application of the modelling method pursues specific goals, and
the effectiveness of the method depends on how skilfully the developer
makes use of the possibilities of modelling.
Regardless of how a particular large system is divided into subsystems,
when designing each of them it is necessary to carry out external
design (macro-design) and internal design
(micro-design). Since at these stages the developer pursues
different goals, the methods and modelling tools used
may differ considerably.
At the macro-design stage, a generalised
model of the system's operating process must be developed, enabling
the developer to obtain answers to questions about the effectiveness of various
control strategies for the object as it interacts with the external environment.
The external design stage can be divided into analysis and synthesis. During
analysis, the control object is studied, a model of the effects of the external
environment is built, the criteria for evaluating effectiveness, the available resources,
and the necessary constraints are determined. The ultimate goal of the analysis stage is to build
a model of the control object for evaluating its characteristics. During synthesis, at
the external design stage, the problems of choosing a control strategy are solved
on the basis of a model of the object being modelled, i.e. the large system.
At the micro-design stage, models are developed with the aim of
creating efficient subsystems. Moreover, the methods and tools of
modelling used depend on precisely which supporting subsystems
are being developed: information, mathematical, technical,
or software subsystems, and so on.
The choice of modelling method and the necessary level of detail of the models
depend significantly on the stage of development of the large system. At the stages
of surveying the control object, for example an industrial enterprise, and
of drafting the technical specification for the design of an automated
control system, the models are mainly descriptive in nature and
pursue the goal of presenting, as fully as possible and in compact form,
the information about the object that the system developer needs.
At the stages of developing the technical and detailed designs of systems, the models
of individual subsystems are made more detailed, and modelling serves to solve
specific design problems, i.e. choosing the optimal variant, according to a
given criterion under given constraints, from the set of
admissible ones. Therefore, at these stages of designing complex
systems, models are mainly used for the purposes of synthesis.
The purpose of modelling at the stage of implementation and operation of
complex systems is to play out possible situations in order to make
well-founded and forward-looking decisions on managing the object.
Modelling (simulation) is also widely used in training and
drilling the personnel of automated control systems,
computing complexes and networks, and information systems in various
fields. In this case, modelling takes the form of business games. The model,
usually implemented on a computer, reproduces the behaviour of the controlled object
and of the external environment, while people, at certain points in time, make
decisions on managing the object.
Automated information and control systems are systems that evolve as
the control object evolves, as new means of control appear, and so on. Therefore, when
forecasting the development of large systems, the role of modelling is very high,
since it is the only way to answer numerous questions about
the paths of further effective development of the system and about choosing the most
optimal one from among them.
The resources of modern information and computing technology make it
possible to set and solve mathematical problems of a complexity
that not long ago seemed unrealisable, for example, the modelling
of large systems.
Historically, the analytical approach to the study of
systems developed first, when the computer was used as a calculator based on analytical
relationships. The analysis of the characteristics of the operating processes of large
systems using only analytical research methods usually runs into
considerable difficulties, leading to the need for
substantial simplification of the models, either at the stage of their construction, or
during the work with the model, which can lead to obtaining unreliable
results.
Therefore, at present, alongside the construction of analytical
models, much attention is being paid to the tasks of evaluating the characteristics of large
systems on the basis of simulation models implemented on modern computers
with high speed and a large amount of operating memory. Moreover,
the promise of simulation modelling as a method for studying
the characteristics of the operating process of large systems increases with
an increase in the speed and operating memory of the computer, with the development
of mathematical support, the improvement of data banks and
peripheral devices for organising interactive modelling systems.
This, in turn, contributes to the emergence of new «purely machine»
methods for solving problems of studying large systems on the basis of organising
simulation experiments with their models. Moreover, the orientation towards
automated workstations based on personal computers for carrying out
experiments with simulation models of large systems makes it possible
to carry out not only the analysis of their characteristics, but also to solve problems
of the structural, algorithmic and parametric synthesis of such systems under
given criteria for evaluating effectiveness and constraints.
The successes achieved in the use of computing technology
for the purposes of modelling often create the illusion that the use
of a modern computer guarantees the possibility of studying a system of any
complexity. At the same time, it is ignored that at the basis of any model
lies a preliminary study, costly in terms of time and material resources,
of the phenomena taking place in the object-original. And how
thoroughly the real phenomena are studied, how correctly
their formalisation and algorithmisation are carried out, ultimately determines the success
of modelling a particular object.
Comments