DEVELOPMENT OF RESOURCE DISTRIBUTION MODEL OF AUTOMATED CONTROL SYSTEM OF SPECIAL PURPOSE IN CONDITIONS OF INSUFFICIENCY OF INFORMATION ON OPERATIONAL DEVELOPMENT

environment and limited computing resources. One of the most problematic places in the distribution of resources of an automated control system is the low quality of planning, distribution and use of resources of an automated system in conditions of insufficient information about the operational situation and the inability to predict the actions of the enemy. This reduces the efficiency of both the system itself and its application. The scientific problem is solved with the help of developing a model for the distribution of system resources with the possible appearance of a lot of perturbations at the input, taking into account the features of the current operational situation in the course of the armed conflict and allows forecasting the state of the automated control system. In the course of the study, the authors of the work used the basic principles of queuing theory, automation theory, the theory of complex technical systems, as well as general scientific methods of cognition, namely analysis and synthesis. The novelty of the proposed model lies in the fact that it allows to justify the decomposition of the system. This allows to present a solution to the vector optimization problem in the binary relations of conflict, assistance and indifference. It also takes into account the operational environment and allows predicting the state of the system taking into account external influences, constructing utility functions and guaranteed payoff, as well as a numerical optimization scheme on this set. The proposed model will improve the efficiency of information processing due to its distribution and rational use of available computing resources. It is advisable to use the research results when planning the configuration of the data transmission system and at the stage of operational control of the resources of these systems.


Introduction
The distribution of resources of the specialpurpose automated control system (ACS) is a function of the sys tem to regulate the use of its resources in the presence of uncertainty about the nature of the conflict using various means of confrontation (destruction of targets). And also because of the inconsistency of the current operational situation with the planned method of using ACS as the most effective for a certain time.
The allocation of resources provides a series of tech niques that help the decisionmaking element (DME) in achieving the best results. In this case, the task of the DME is to find ways to respond flexibly to changes in the operational environment in order to minimize the impact of the current situation on the automated control system using the means of distributing its resources.
However, existing approaches to the allocation of resources do not satisfy the requirements that apply to them, namely: -great computational complexity [1,2]; -need to know complete information about the state of the system and the actions of the enemy [3,4]; -inability to predict the actions of the enemy on the state [5][6][7].
In this regard, it is necessary to develop a model for the distribution of ACS resources in conditions of insuf ficient information on the operational situation, taking into account the essential features of the course of the armed conflict in a specific situation. So, the object of research is an automated control system for special purposes in the

Methods of research
The resulting control u, which affects the control object, is formed by two components, such as software (plan x), and corrective (operational control c) [1,2]. This type of control can be represented in different ways: as the sum of two components, or some other given by their function: The operating party must ensure that all conditions of admissibility of the resulting control are met, can be briefly written as follows: where U -a given set in the functional space of controls, depending on perturbations ε. In addition, the operating side is trying to maximize the quality criterion of controls, which is also affected by perturbations: Records (2) Then, the inverse influence of the control on the set of expected perturbations Ξ 0 is taken into account. Based on a priori information I Ξ ( ) about perturbations, the ope rating side determines the plan x in advance: The next step will be to consider two options for a priori awareness [7,8]: -only a lot of future perturbations Ξ 0 are known, which is set, for example, by the limiting values of per turbations; -only a distribution function of perturbations µ Ξ ( ) are known, that is, the probability with which perturbations can fall into any subset Ξ of the original set Ξ 0 . Operational control is formed after planning, in the process of functioning, using information i ε ( ) about the perturbation. This makes it possible to compensate for some unwanted effects: By operational control algorithm is meant a complete list of actions for each of the existing implementations ε ∈Ξ 0 . The construction of such algorithms is not considered in this case, therefore, let's consider the operator (5) as given. When the information at the given moment of time is complete, that is i ε ε ( )= , it is possible to analytically create the optimal operational control algorithm in some solved quasistatic problems [9,10].
To complete the operational task, a certain set of re sources is needed, part of which is in the ACS, and the other in the higher system (supersystem), the volume of centralized replenishment of the ACS x resources is planned to be operated by the party -ACS operator in advance. The ACS operator can't exert any influence on the volume and availability of resources of a higher super system ε, and also, when planning, the value ε is unknown, that is, it is in the category of perturbations.
The total amount of resources x + ε consists of two parts, one is transferred to the ACS and has a single capacity, the other may be, depending on the operational situation, in a passive reserve with a capacity r, or be ready for use by a higher system [3]. On the other hand, the total amount of resources x + ε should be enough for the ACS in the conditions of insufficient information on the development of the operational situation, together with a higher system, to be able to carry out the necessary opposition to the enemy. To do this, it is necessary to provide an initial ith share of the maximum (single) complex of ACS resources 0 1 < < ( ) υ . The missing amount of resources can be filled with the re distribution of ACS resources or with the actions of a higher system. The operational control effect in this consists in the redistribution of ACS resources to perform a partial task.
The operational control effect in this consists in the redistribution of ACS resources to perform a partial task. Positive values of the control action correspond to the replenishment of the stock necessary to achieve a partial goal, negative -the reduction of the necessary stock of the resource [2].
The resulting control (1) is equal to the sum of the planned x and operational in terms of: which in the current setting are not vector functions of time, but simply scalar parameters that the DME of the operating side selects. The set (2) of possible resulting controls is determined by the ACS resource reserves reduced to unity, the mini mum allowable amount υ, and also the size r of the passive reserve. Such a set significantly depends on perturbationsthe quantities ε and u are taken under conditions of ad missibility of control on equal rights: u U x y x y r x y The control quality is assessed by the quality of the task, which linearly depends on all three types of resources: centralized x (ACS resources), decentralized ε (supersys tem resources) and compensating у (redistributed) [1,3]: where g q ≥ ≥ ≥ 1 0. Different values of the coefficients g, q reflect the mis match of the confrontational situation, which for sim plicity is considered already known at the planning stage. Ac cording to the above, operational control y is carried out according to reliable information about perturbations ε and plan x. So, from the admissible set (7), it is possible to choose the value y, maximizing the criterion (8) Expression (9) is an operational control algorithm (5), according to which the compensating amount of resources is always equal to the maximum possible: y y u = . This either ensures the full use of ACS resources, if y E r = − − ≤ 1 ε , or completely exhausts the supply of resources of the supersystem, if y r x = ≤ − − 1 ε. The lower bound y u is in volved only in creating conditions for the admissibility of algorithm Y; close relation (9). The admissibility of operational control is guaranteed by choosing the plan x.
Given the planning of the operational control algo rithm Y, condition (2) for the admissibility of the re sulting control must be fulfilled only by choosing the plan x. This provides an adequate supply of regulatory resources.
Conditions (2), reflected in the space of plans x, can be divided into three types: 1. Conditions that are independent of the perturbation, as well as the conditions of integer or, in the general case, discreteness of some components of the plan vector x: . for (10) Expression (10) shows the process of constructing a set X 1 . This is the union of a finite number of indepen dent subsets X n 1 , each of which is possibly compact and depends on the number n. Allowed and purely discrete options when the subsets X n 1 are finite or pairwise. 2. Conditions that depend on the perturbation, but according to their purpose, must be realized for all a priori possible implementations of the perturbation: 3. Conditions depend on the perturbation: In (11)  can be assigned to expression (10), which has no pertur bation, due to the equivalence of two inequalities: in the absence of dependence on the perturbation ε in terms of х. As a result, the condition for the simultaneous execution of expressions (10) and (11) can be reflected in a simpler form: For the convenience of further reasoning, the perturba tion vector ε is divided into two subgroups -continu ous η ( ) and discrete χ ( ):

Research results and discussion
The separation of all possible perturbations into dis joint subsets of favorable and unfavorable perturbations is performed for a fixed set of plans x and a fixed lower bound for the implementation J  of the quality criterion for the resulting control.
It is clear that for the subsequent consideration, plans that correspond to the conditions (17) of the perturba tions, do not include perturbations are important.
In (18) and in Fig. 1 is indicated by: ε u0 -upper a priori estimate of perturbations; ε u1 -maximum perturbation in terms of resources allowed by the supersystem; ε l 0 -lower a priori estimation of perturbations; ε l1 -minimal perturbation, permissible on condition of loading; ε l 2 -minimal perturbation, provides the expected re sult c with the possibility of a free set of resources; ε l 3 -minimum perturbation, provides the expected result c in the presence of the remaining resources. With increasing intensity of the confrontation c, the lower bounds ε l 2 and ε l 3 , are raised, while others remain constant. In accordance with this, the range of favorable perturbations (18) narrows, degenerating in a certain sense value c into an empty set. This is a general property of the sets of favorable perturbations (16), (17).
Next, problematic questions arise related to the assess ment of the influence of perturbations on the effectiveness of counteraction, which allows the use of the research results in models of supporting decisionmaking.

Conclusions
During the study, the authors developed a model for the distribution of ACS resources under conditions of insuf ficient information on the development of the operational environment. The novelty of the proposed model is that it: -allows to justify the decomposition of the system, allows to imagine a solution to the vector optimiza tion problem in the binary relations of the conflict, assistance and indifference; -takes into account the operational environment; -allows to predict the status of the system, taking into account external influences; -allows to build utility functions and guaranteed payoff, as well as a numerical optimization scheme on this set. The proposed model will improve the efficiency of in formation processing due to its distribution and rational use of available computing resources.
It is advisable to use the results of the study when planning the configuration of the data transmission system and at the stage of operational control of the resources of these systems.