Hierarchical management of large systems. Hierarchical management structure of an organization: features, principles, types

One of the main means of overcoming the organized complexity of a system is decomposition, i.e. dividing the system into parts (the so-called "black boxes") and organizing these parts into a hierarchical system. The system is divided into subordinate parts so that each part contains objects that are most closely related to each other. Consequently, the system is divided according to weak connections.

Decomposition is a conventional technique that ultimately allows one to evaluate the degree of complexity of an object and reduce it to some finite elements, the analysis of which can be performed using known methods. We will assume that element - This is a part of the system, the further division of which leads to disruption of the functional connections of the element and the obtaining of properties of the selected set that are not adequate to the properties of the element as a whole.

The benefit of using black boxes is that the user only needs to know the input and output of the black box and its purpose, i.e. the function performed, without going into the principles of operation and the algorithms used. In everyday life, we often come across “black boxes” and readily use them. For example, we use a printer to prepare documents without knowing how it transcodes and prints information. We can replace the printer with another one if it breaks down or with a more modern one, without being technical support specialists. The idea of organizing “black boxes” into hierarchical structures was taken by man from nature. All complex systems of the Universe are organized in hierarchies. And the Universe itself includes galaxies, star systems, planets, etc.

Hierarchical system

If a set of elements is combined into a system according to a certain characteristic, then it is always possible to introduce some additional characteristics to divide this set into subsets, thereby separating its component parts from the system - subsystems. The possibility of repeated division of a system into subsystems leads to the fact that any system contains a number of subsystems obtained by separation from the original system. In turn, these subsystems consist of smaller subsystems, etc.

Subsystems obtained by separation from one source system are classified as subsystems of the same level or rank. With further division we obtain subsystems of a lower level. This division is called hierarchy(division of positions into higher and lower, the order of subordination of lower-ranking persons to higher-ranking ones, etc.). The same system can be divided into subsystems in different ways - this depends on the chosen rules for combining elements into subsystems. The best, obviously, will be a set of rules that provides the system as a whole with the most effective achievement of the goal.

When dividing a system into subsystems, you should remember the rules for such division:

each subsystem must implement a single function of the system;

the function allocated to the subsystem should be easily understood regardless of the complexity of its implementation;

communication between subsystems should be introduced only if there is a connection between the corresponding functions of the system;

connections between subsystems should be as simple as possible.

The number of levels and the number of subsystems of each level may be different. However, one important rule must always be followed: subsystems directly included in one higher-level system, acting together, must perform all the functions of the system they are part of.

Management of any organization that produces goods or provides services is built on a hierarchical principle. Activities to create goods and services occur in all organizations. Production - This is the creation of goods and the provision of services by transforming the input of the system (necessary resources of all types) into its output (finished goods and services). In manufacturing firms, product creation activities are usually obvious. Its result is specific goods (for example, machines or airplanes). In other organizations. who do not create physical goods, production functions may be less obvious, hidden from the public and each of the buyers. For example, this is an activity that is carried out in a bank, an airline office or a college. The activities of such companies are called service. Operations managers make the decisions necessary to transform resources into goods and services.

In a hierarchical control system, any subsystem of a certain level is subordinate to a higher level subsystem, of which it is a part and is controlled by it. For control systems, division of the system is possible until the subsystem obtained during the next division ceases to perform control functions. From this point of view, the control system of the lowest hierarchical level are those subsystems that directly control specific tools, mechanisms, devices or technological processes. A control system of any level other than the lowest one always controls technological processes not directly, but through subsystems of intermediate, lower levels.

An important principle for constructing an enterprise management system is to consider the enterprise as a system with a multi-level (hierarchical) structure (Fig. 1.2). From links located at a higher level, there is a flow of control actions, and information about the current state of a control object at a lower level is supplied to links at a higher level. Considering a kind of “tree” of management, it can be noted that the advantage of a hierarchical management structure is that the solution of management problems is possible on the basis of local decisions made at the appropriate levels of the management hierarchy.

Rice. 1.2. Hierarchical enterprise management systems

The lower level of management is a source of information for making management decisions at a higher level. If we consider the flow of information from level to level, then the amount of information, expressed in the number of symbols, decreases with increasing level, but at the same time its semantic content increases.

At the current level of development of society, scientific and technological progress in the field of material production and management systems provides the opportunity to concentrate and centralize significant financial, material and other resources. These opportunities are realized in industrialized countries in the form of the creation of international associations (for example, the European Union, which unites a number of European countries; subsidiaries, branches and enterprises of large concerns in many countries of the world, etc.). Advantage centralization is the ability to direct large resources to implement solutions, which makes it possible to solve complex problems that require large capital investments. In a centralized system, it is relatively easy to ensure coordinated, coordinated activities of subsystems aimed at achieving common goals. Losses in individual parts of the system are compensated by the results of the work of its other parts. A multi-level centralized system has great survivability due to the prompt redistribution of functions and resources. It is no coincidence that the principle of centralization is strictly observed in the armies of all times and peoples.

However, centralization in large-scale systems has its drawbacks. Multi-levelness and the associated repeated transfer of information from level to level causes delays that reduce the efficiency of assessing the situation and implementing management decisions, leading to distortions both in the process of transmitting information and during its processing at intermediate levels. In a number of cases, the desire of subsystems for independence comes into conflict with the principle of centralization. In multi-level centralized organizational and administrative management systems, as a rule, there are elements decentralization.

With a rational combination of elements of centralization and decentralization, information flows in the system must be organized in such a way that information is used mainly at the level where it occurs, that is, one must strive for minimal data transfer between levels of the system. In decentralized single-level systems, there is always a higher level of efficiency both when collecting information about the state of the managed system, assessing the situation, and when implementing decisions made. Thanks to operational control over the reaction to control inputs, deviations from the chosen trajectory of movement towards the goal are reduced.

The degree of centralization of the system, which is determined on the basis of establishing the ratio of the weighted volumes of tasks solved at adjacent levels, serves in a certain sense as a measure of the division of powers between levels. The shift of the bulk of decisions towards a higher level, i.e., an increase in the degree of centralization, is usually identified with an increase in the controllability of subsystems. It requires, as a rule, improved information processing at the upper levels of the management hierarchy. An increase in the degree of decentralization corresponds to an increase in the independence of subsystems and a decrease in the amount of information processed by the upper levels.

Typically, senior managers of multi-level systems develop strategic decisions, for example, how many car models each of the company's plants should produce. They should not decide on the size and quantity of each model produced at each plant. This applies to the level tactical decisions, which are accepted by plant middle managers. The factory manager must decide how much to produce and sell, how much to store in finished goods inventory (seasonal demand), and how many workers to hire or fire. Operational Decision Making carried out at the production level by shop managers who determine detailed planning and production. This hierarchical approach, which must include feedback, may not provide the optimal solution, but it does allow for better and more timely control of the production process.

The structure of management systems in the national economy is built on a sectoral or territorial principle. Industry principle is used in cases where we are talking about complex, specific types of production, design and construction, the development and implementation of scientific research in production of a certain type. By territorial principle State administrative bodies were built.

Any modern organization, be it a commercial company, an industrial enterprise, or in order to achieve its goals, must have a clear and clear management structure. If we start from the definition, then the management system of an organization is a set of interconnected and interdependent units and individual individuals who fill certain positions, who are not only in the “superior-subordinate” position, but also directly influence development of this organization.

The system is not created at once; it is a rather lengthy process that includes the following main stages:

At the first stage, the leadership core determines what kind of structure will be created: a hierarchical structure, functional or direct subordination.
The second stage includes the creation and empowerment of the main structural components, such as the management apparatus itself, programs, and divisions.
Finally, at the third stage, the final redistribution of power, duties and responsibilities occurs. At the same time, it is advisable to consolidate all these powers in the form of regulations on certain departments and job descriptions.

Although many types of management structures are known today, one of the most popular is the hierarchical management structure. It was theoretically substantiated and experimentally tested at the beginning of the twentieth century by an American sociologist. Subsequently, most scientists were mainly engaged in finding more and more new ones of this system.

The hierarchical management system is based on the following principles:

The entire management system is a pyramid, each lower level of which is subordinate to and controlled by a higher one.
A hierarchical structure implies a clear division of authority between levels. In this case, the higher level bears higher responsibility compared to the lower level.
Labor in any organization that is managed according to a hierarchical principle must be clearly divided among its workers, who specialize only within the functions they perform.
Any activity in an institution with a hierarchical management structure must be standardized and formalized. In this way, better coordination of workers’ activities will be achieved and their level of controllability will increase.
Hiring should only be made in accordance with the employee's requirements. At the same time, in addition to professional qualities, it is necessary to pay attention to how well this employee is managed and how ready he is for the role of manager.

The hierarchical structure implies that all employees of the organization can be classified into one of three main groups - managers, specialists and performers. Moreover, since all organizations are very similar in their management type, managers can use the experience of their colleagues to make their management structure more optimal.

The main types of hierarchical management structures should be considered a linear structure, where all the main threads are concentrated in the hands of the boss, functional, when each division of the organization is engaged in performing a specific function, as well as a mixed type of management, where, along with the linear apparatus, there is a branched hierarchy of various functional groups.

When dividing a system into subsystems, you should remember the rules for such division:

Each subsystem must implement a single function of the system;

· the function allocated to the subsystem should be easily understood regardless of the complexity of its implementation;

· communication between subsystems should be introduced only if there is a connection between the corresponding functions of the system;

· connections between subsystems should be simple (as far as possible).

Rice. 1.2. Hierarchical enterprise management systems

A large system, as briefly described in Chapter 1, is a complex system made up of many components or smaller subsystems that perform functions, share resources, and are governed by interrelated goals and constraints (Machmoud, 1977; Jamshidi, 1983). Although the interaction of subsystems can be organized in various forms, one of the well-known ones is hierarchical, which is natural for economics, management, enterprise management, and mixed industries such as robotics, oil, steel and paper production. In these hierarchical structures, subsystems are located at levels with varying degrees of hierarchy. A subsystem at any level controls or coordinates subsystems located at a level below it, and, in turn, is controlled or coordinated by a subsystem located at a level above. Figure 4.1 shows a typical hierarchical (multi-level) system. The top level of management, sometimes called the supreme coordinator, can be compared to the board of directors of a corporation, while the other levels can be compared to the president, vice president, directors, etc. The lowest level could be, for example, a plant manager, a store director, etc. whereas the larger system itself is the corporation. Although the representation of hierarchical structure seems quite natural, its precise behavior is not yet fully understood due to the fact that little research has been done in the field of large systems (March and Simon, 1958). Mesarovic et al. (1970) presented one of the earliest formal quantitative approaches to a hierarchical (multilevel) system. Since then, much work has been done in this area (Schoeffler and Lasdon, 1966; Benveniste et al., 1976; Smith and Sage, 1973; Geoffrion, 1970; Schoeffler, 1971; Pearson, 1971; Cohen and Jolland, 1976; Sandell et al., 1978; Singh, 1980; Jamshidi, 1983; Huang and Shao, 1994a,b). The interested reader can find relatively comprehensive information on the management of multilevel systems and their application in Mahmoud (1977).

This section describes the concept of hierarchy, the properties and types of hierarchical processes, and presents some reasons for their existence. A full evaluation of hierarchical methods is presented in Section 4.6.

The following are the main properties of hierarchical systems, although they are not generally accepted:

1. The hierarchical system consists of control blocks, which are organized according to the pyramid principle.

2. The system has an overall goal, which may or may not coincide with the goals of the individual components of the system.

3. Different levels of the system hierarchy repeatedly exchange information with each other (usually vertically).

4. As the level increases, the time range also increases, that is, the components of the lower levels are faster than the components of the upper ones.

In hierarchical (multi-level systems) three main structures can be distinguished, depending on the parameters of the model, the desired variables, behavior and environment, variability and the existence of many mutually exclusive goals and objectives.

1. Multilayer hierarchical structure. In this multi-level structure, the levels are called layers. Lower-level subsystems provide a more accurate description of the larger system than upper-level subsystems.

2. Multilayer hierarchical structure. This structure is a result of the complexity of the regulatory process. Management tasks are distributed vertically as shown in Figure 4.2 (Singh and Titli, 1978). In a multilayer system, which is shown in the figure, regulation (at the first level) is direct control, followed by optimization (calculation of control points of regulators), adaptation (direct adaptation of the control law and control model) and self-organization (choice of model and control as a function environmental parameters).

3. Multi-link hierarchical system. This is the most common of all three structures; it consists of several subsystems, which are located at levels in such a way that each level (as described above) can control the subsystems of the lower level, and is controlled by the subsystems of the upper levels. This structure, depicted in Figure 4.1, takes into account the mutually exclusive goals and objectives of the various sublevels. In other words, higher-level stages achieve mutually exclusive goals by weakening the interaction between lower-level stages. The distribution of the control task of this structure is shown in Figure 4.2 and, unlike a multilayer structure, it is horizontal.

In addition to the vertical and horizontal distribution of management tasks, there is a third method - temporary or functional distribution. This distribution, which gives the subsystems a functional optimization of the problem, consists of decomposing the problem into a finite number of simple optimization problems at the lower level and as a result gives a considerable reduction in calculations. This scheme was used for hierarchical control of discrete systems by Jamshidi (1983).

The remainder of this chapter discusses how hierarchical systems can be managed effectively using processes known as decomposition and reconciliation. These two processes are shown in Figure 4.3. In summary, the definition of hierarchical control is: (a) decomposition - dividing a system into many subsystems, and (b) coordinating the work of these subsystems until optimal control of the entire system is achieved (through a multi-level iterative algorithm).

Section 4.2 describes the possibility of applying reconciliation to hierarchical systems. Section 4.3 covers open-loop control. Section 4.4 is devoted to closed-loop control; it also provides definitions of “interaction prediction” and the structural perturbation method. Section 4.5 describes hierarchical control based on Taylor and Chebyshev series expansion. The control problem is solved by linear algebraic equations. Examples show various solution methods. Optimization of linear and nonlinear hierarchical systems is described in Chapter 6. Section 4.6 contains further development of hierarchical control methods.

Hierarchical system

Parameter name	Meaning
Article topic:	Hierarchical system
Rubric (thematic category)	Electronics

Any system contains a number of subsystems obtained by separation from the original system. In turn, these subsystems consist of smaller subsystems, etc.

Each subsystem must implement a single function of the system;

· the function allocated to the subsystem should be easily understood regardless of the complexity of its implementation;

· communication between subsystems should be introduced only if there is a connection between the corresponding functions of the system;

· connections between subsystems should be simple (as far as possible).

The number of levels, the number of subsystems of each level should be different. However, it is always necessary to observe one important rule: subsystems that are directly included in one higher-level system, acting together, must perform all the functions of the system to which they are included.

Management of any organization that produces goods or provides services is built on a hierarchical principle. Activities to create goods and services occur in all organizations. Production - This is the creation of goods and the provision of services by transforming the input of the system (necessary resources of all types) into its output (finished goods and services). In manufacturing firms, product creation activities are usually obvious. Its result is specific goods (for example, machines or airplanes). In other organizations that do not create physical goods, production functions may be less obvious, hidden from the public and individual customers. For example, this is an activity that is carried out in a bank, office, airline or college. The activities of such companies are called service. Operations managers make the decisions necessary to transform resources into goods and services.

In a hierarchical control system, any subsystem of a certain level is subordinate to the higher-level subsystem of which it is a part and is controlled by it. For control systems, division of the system is possible until the subsystem obtained during the next division ceases to perform control functions. From this point of view, the control system of the lowest hierarchical level are those subsystems that directly control specific tools, mechanisms, devices or technological processes. A control system of any level other than the lowest one always controls technological processes not directly, but through subsystems of intermediate, lower levels.

An important principle for constructing an enterprise management system is to consider the enterprise as a system with a multi-level (hierarchical) structure (Fig. 1.1). From links located at a higher level, there is a flow of control actions, and information about the current state of a control object at a lower level is supplied to links at a higher level. Considering a kind of “management tree,” it can be noted that the advantage of a hierarchical management structure is that the solution of management problems is possible on the basis of local decisions made at the appropriate levels of the management hierarchy.

Hierarchical system - concept and types. Classification and features of the category "Hierarchical system" 2015, 2017-2018.