ON ONE APPROACH TO DESIGNING AUTOMATED CONTROL SYSTEMS

G. F. Mazurek, Yu. V. Karyakin

(Presented by the scientific and technical seminar of the Department of Automation and Telemechanics)

The broad scope of design and survey work on the creation of automated control systems (ACS) for organizations of various kinds requires sufficiently clear and well-developed methodological materials on these issues. In our opinion, the methodological guidelines on surveying organizations and conducting design work that have been published to date have the following shortcomings:

1) survey methods, as a rule, are not linked with models for analyzing the organization and synthesizing the ACS;

2) most methods for surveying organizations orient developers toward studying information flows in the control system, leaving in the background other aspects of its construction and functioning; moreover, information is studied at the syntactic level and more rarely at the semantic level, although the value of information in control tasks is a most important characteristic;

3) in most methods for studying organizations, the survey results are presented in the form of various information diagrams; such a representation is suitable for control systems of small dimension, but it does not provide a sufficiently high level of analysis of the survey results and practically precludes machine processing of the materials;

4) in the overwhelming majority of methods, analysis of the survey results is carried out at the substantive level, i.e., on the basis of speculative considerations concerning the rationalization of flow diagrams; in this case the rationalized diagram includes the information flows of the existing control system without its substantial alteration;

5) the methods do not analyze the influence of the improved information system on other characteristics of the control system;

6) most models for analyzing the control system and synthesizing the ACS are of a particular nature, which makes it difficult to link them into a single complex of design work. A detailed analysis that made it possible to identify the indicated shortcomings is presented in [1].

These shortcomings are caused by both objective and subjective factors. Let us consider individual methodological aspects of creating an ACS that are connected with the listed shortcomings of existing methods.

To conduct an analysis of the control system with the aim of creating an ACS, we propose considering the organization in four different aspects and, in doing so, distinguishing the following structures: functional, organizational, economic, and informational. In constructing the structures, graph models are used. There is no doubt that any pair of the named structures is characterized by mutual reciprocal influence; therefore, when examining and analyzing them, they cannot be considered in isolation from one another.

Consideration of these structures from the point of view of the sequence of their construction in the process of creating a model of the control system made it possible to represent the genetic interrelation of the structures in the form of a matrix (Table 1). The ones in the cells of the matrix indicate that the structure named in the row is the basis for constructing the structure named in the column. Identifying the genetic interrelation of an organization’s structures creates prerequisites for a deeper and more comprehensive analysis of the control system and indicates the choice of methods for the sequential synthesis of ACS structures. In Table 1, moreover, the objects of the organization subject to examination and the methods for studying them in order to construct the corresponding structures are listed, as are certain tasks of analyzing each of them and also individual tasks of synthesizing ACS structures.

The approach set forth also makes it possible, already at the stage of examining the organization, to carry out a clear separation of the objects and methods of examination for constructing the corresponding structures.

Thus, although the examination of an organization for the purpose of rationalizing management or designing an ACS is a unified process, methodologically it is divided into several different directions.

The first direction is subordinated to the goal of identifying and constructing the functional and organizational structures of the organization. The objects of study for constructing these structures are the same, and the methods of examination coincide (items 1, 2, 4 in the left-hand column of Table 1).

The second direction has as its goal the construction of the economic and informational structures of the organization. In this case, all the examination methods listed in the left-hand column of the table are used. When constructing the informational structure, the task is taken as the unit of observation of information (as an element of the informational structure). In a more in-depth study of the informational structure, it is divided into two independent directions: the study of control operations and the description of information flows.

As was noted at the beginning of the article, methodologies for examining organizations are, as a rule, not linked to models for their analysis. Moreover, often in examination methodologies the models for analyzing organizations are not considered at all. A more rational approach to examining an organization appears to be one in which the examination methodology is created on the basis of developed models of analysis. In this case, the examination materials must satisfy the analysis models in terms of completeness of information. This point of view is reflected in [2].

Table 2 presents one of the possible schemes of the process of developing, in the first stage, an ACS. Here, blocks numbered 1–8 represent the results of examining the organization by its functions, tasks, information flows, etc. In practice, this is the completion of specially developed cards for examining the organization.

Blocks numbered 9–17 represent models for analyzing the organization and developing the ACS.

Blocks numbered 18–21 are the results of analyzing the organization and synthesizing the ACS.

Table 1

Key to Table 2:

— genetic interconnection of structures with minimal composition of indicators.

→ survey—investigation—modeling—
analysis
← analysis—synthesis

TASKS OF ACS SYNTHESIS

1. Development of an economic structure with a minimal composition of indicators.
2. Development of a functional and informational structure of the technical-and-economic control system optimal in terms of cost.
3. Development of a functional and informational structure of operational control optimal in terms of speed of response.
4. Selection of ACS technical means.
5. Development of the organizational structure.

Table 2

Input

1. Structure of management of the organization
2. Permissible instructions and regulations
3. Description of the tasks solved by the operators
4. Description of information flows
5. Normative economic indicators
6. Description of material flows
7. Management method
8. Special requirements

Analysis–Synthesis

1. Preparation of initial data for analysis
2. Analysis of data for selecting the technical means of the ACS
3. Analysis of tasks and their ordering
4. Analysis of data for substantiating economic efficiency
5. Selection of the technical means of the ACS
6. Synthesis of the information system for data processing (ISDP)
7. Substantiation of economic efficiency
8. Analysis–synthesis of the management structure
9. Analysis–synthesis of information flows

Output

1. Recommendations on the technical means of the ACS
2. Recommendations on the management structure
3. Recommendations on the organization of information flows
4. Expected economic effect

The forms of survey cards can be developed so that they unambiguously reflect the selected model of analysis and contain not only links but also numerical characteristics of the elements of the system. The most convenient for analyzing an organization are mathematical models of graph theory. In this case, the nodes of the graph may be functions, tasks, operators, economic indicators, technical means, etc., and the arcs may be the links between them. When such a model of the survey card is chosen, in essence, the adjacency matrices of the corresponding graphs are specified. This makes it possible to punch the survey cards and carry out machine processing of the data.

References

1. G. F. Mazurek. Analysis and classification of methods for studying information flows in organizational management systems. Article in the present collection.

2. General sectoral guiding methodological materials on the creation of automated enterprise management systems (ASUP). TsNIITU, Minsk, 1972.