Formulation of the problem
The process of enterprise management consists of the development, adoption and implementation of strategic and tactical management decisions. Limited human resources and a constant desire to reduce costs have led to the creation of systems that can take into account various aspects that can influence the choice of one option or another in the decision-making process, and also determine which options are more preferable.
The easing of the management function in planning the execution of orders and organizing their implementation is a large-scale task, which includes several sub-tasks.
DSS is a collection of intelligent information applications and tools that are used to manipulate, analyze, and present the results of such analysis to the end user. Modern DSS allows to provide for the degree of influence of the decisions taken on the further development of the business. Thus, the DSSP unites on a common basis approaches that are typical for such areas: decision making, knowledge acquisition and presentation, construction of human-machine (interactive) systems.
Developed DSS is necessary to facilitate the process of organization of production and will solve two tasks:
1. Evaluation of the feasibility of the order under consideration based on the situation at the enterprise, as well as order parameters
2. The choice of the optimal sequence of execution of the order series under consideration proceeding from the situation at the enterprise, as well as the order parameters
Evaluation of the feasibility of the order in question
Solving this problem will make it easier for management to choose between whether to reject a given order and not to undertake it, or vice versa, to take it. Thus, it can be concluded that the result of the work of this function developed DSS should be advice to the leadership of two possible:
1. Execution of this order is expedient
2. The execution of this order is inappropriate
Analyzing the parameters of the order, as well as the current situation in the company, you can determine the degree of utility of the order for the enterprise. The degree of utility of the order is the determining criterion, which will determine the feasibility of this order.
The choice of the optimal sequence of the order in question
The calculation of the optimal sequence is the task of ranking. The solution of the task of ordering orders takes place in 2 stages:
1. Evaluation of the utility of each order
2. Order Ranking
Analyzing the parameters of each of the orders, as well as the current situation in the enterprise, you can determine the degree of its usefulness for a given enterprise, which will determine its priority within the order fulfillment queue at the enterprise and rank orders, forming the optimal sequence for the company to execute them.
The choice of the optimal sequence of the order in question
The work of the enterprise is influenced by a number of factors that need to be analyzed in order to take into account their influence in the developed DSS. The following internal factors were identified:
1. Customer Reliability (Function N)
2. The degree of sufficiency of materials for this order in the warehouse (function M)
3. The degree of workload of workers required to perform this order (function Z)
4. Care of employees needed to fulfill this order for a sick leave (function B)
5. Profitability (function P)
The work of the enterprise is influenced by a number of factors that need to be analyzed in order to take into account their influence in the developed DSS. The following internal factors were identified:
1. Power supply stability (SE function)
2. Stability of water supply (SV function)
3. Stability of the general political situation in the region (function SP)
4. Stability of the supply of materials required for the production process (SPM function)
Model of decision support system
Based on all listed production factors, the degree of utility of each order for the enterprise is determined. By such a characteristic as the utility of an order is understood the aggregate of two calculated coefficients for each of the orders:
1. Coefficient of comfort
2. Coefficient of profitability
Since the usefulness of the order is determined by the above production factors, accordingly, the described coefficients for each order depend on the factors of production. The coefficient of comfort (KoefU) depends on the degree of sufficiency of materials, the degree of workload of workers, on the factor of care of workers to the sick-list, the stability of electricity supply, the stability of water supply, stability of the general political situation in the region, the stability of supplies of materials necessary for the production process.
KoefU=F(M, Z, B, SE, SV, SP, SPM);
The coefficient of profitability (KoefP) depends on such factors as profitability and reliability of the customer:
KoefP=F(N, P);
Because of the unequal factors, it was decided to evaluate their importance in order to take this into account when learning a neural network. For this purpose the hierarchy analysis method (HAM) was chosen. This method allows us to assess the significance of each factor within the framework of the task.
The DSS should evaluate the feasibility of order fulfillment and the selection of the optimal sequence of order fulfillment. In each of these functions it is necessary to determine the convenience and profitability of each order. To realize this task, it was decided to use a neural network, which will be trained on the data of previous orders of a really functioning enterprise for manufacturing printed circuit boards. Data includes sets of factors and assessments of convenience and utility for each order within these factors. The inputs of the neural network are given production factors, and at the output we get the coefficients of profitability and convenience (Figure 1).
Figure 1 - Calculating the coefficients of convenience and profitability
The data that is necessary for training the neural network is stored in the database. These data include the values of production factors and the corresponding expert assessments, as well as the expert values of the valuation of each production factor in terms of its significance. The database for training the neural will be filled with data provided by a private enterprise engaged in the manufacture of printed circuit boards.
In turn, the data that will be needed in the work of the decision support system at the enterprise, namely the values of the production factors of the orders under consideration or already considered, in other words, all the data participating in the DSS work is stored in a special database, which will provide information support for the decision support system.
In order to choose the best sequence of orders, all alternatives must be considered. Each alternative is one specific order execution sequence. The usefulness of each alternative is determined by the sum of the utilities of all orders that are part of the sequence. Also it should be noted that the usefulness of the order is not the same in different alternatives, because the factor of convenience includes factors that depend on the order of execution of orders. This means that each calculation of the utility of each alternative should recalculate the usefulness of each order. In order to determine the optimal sequence of order fulfillment, it is necessary to consider all possible alternatives, having found within each framework a common coefficient of convenience and profitability. After that it is necessary to find the most optimal alternative. For this process, the method of minimum distances is chosen [4]. Each alternative will have 2 characteristics according to 2 criteria - convenience and profitability, respectively. In this method, each alternative is considered as a point in 2-dimensional space, and its characteristics for each criterion coordinate along the corresponding axis (Fig. 2). In addition to the alternative points (blue dots in Figure 2), there will be another point on the formed 2-D coordinate system - the Ideal (the green dot in Figure 2). This point has coordinates corresponding to ideal values for each of the criteria, hence, it is absolutely optimal within the framework of this process.
Figure 2 - Diagram of the functioning of the method of minimum distances
Thus, the distance between two points is equivalent to the degree of difference between the two alternatives simultaneously by all the criteria. This means that by calculating the difference of each alternative point from the Ideal point, one can determine the order of optimality of alternatives according to the following logic: the less the alternative is different from the ideal, the more optimal the alternative, respectively, the smaller the distance between the alternative point and the ideal point (purple dotted line 2), the alternative is optimal. Hence, the distance between a given alternative point and the Ideal point can be considered the rank of this alternative and is calculated by the following formula [4]:
R(alernative)=((X(i)(ideal)-X(i)(alternative))^2+(X(i+1)(ideal)-X(i+1)(alternative))^2+...
+(X(N)(ideal)-X(N)(alternative))^2)^1/2
Where, alternative is the number of the current alternative point under consideration, ideal-point Ideal, i-number of the axis, X (i) (alternative) - alternative-alternative coordinate, and R (alernative) - distance from the current alternative point alternative to the ideal point (the rank of the alternative).
The alternative with the lowest rank and will be the most optimal order execution sequence.
In the case of the function of evaluating the feasibility of executing one order, an analysis of the obtained convenience and utility factors is made by comparing them with predetermined constraints determined by the management of the enterprise on the basis of personal preferences and corporate rules. If the resulting coefficients are above the minimum values (constraints), then the fulfillment of this order is advisable, otherwise - no. Thus, the obtained coefficients of convenience and profitability are the criteria for assessing the feasibility of the order for this enterprise:
1. - IF KoefU> = A And KoefP> = B THEN the order should be executed;
2. - In other cases - no.
Here A is the minimum value of the convenience ratio, determined by the management of the enterprise on the basis of personal preferences and corporate rules, and B is the minimum value of the profitability coefficient
In the case of the function of choosing the optimal sequence of orders, the pairs of coefficients obtained for each alternative order of orders will be used as expert estimates for the ordering of alternative orders using the minimum distance method described above. As a result, a sequence of orders will be obtained, which will be optimal.
Conclusions
A model of a decision support system for a small enterprise has been developed that will help improve the organization of production by facilitating the management function, reducing the risk of making an incorrect production decision, which will reduce losses and increase the profit of the enterprise. The developed model of the DSS includes tools for analyzing and collecting data necessary for making decisions that will allow providing information support to DSS, the decision-making apparatus itself that directly performs the role of a mechanism that facilitates the governing function and gives recommendations on the expediency of fulfilling an order and constructing an optimal order execution sequences, as well as a user interface that allows you to enter order parameters and get recommendations on the feasibility of performing or to determine the optimal sequence for the fulfillment of orders received by the enterprise.
Список литературы
1. Сергей Гриняев. Нечеткая логика в системах управления [Текст]/ Сергей Гриняев: - Журнал «Компьютерра», Издательский дом «КОМПЬЮТЕРРА», 2002.
2. И. В. Матях, Е.О. Савкова, Корреляционный анализ методов определения весовых коэффициентов значимости для системы оценки социально-экономических показателей развития предприятия[Текст]/ И. В. Матях, Е.О. Савкова - ИНФОРМАТИКА И КИБЕРНЕТИКА №2, До-нецк ДонНТУ, 2015.
3. Fundamental research [Electronic resourse] / Интернет-ресурс. Режим доступа: https://fundamental-research.ru/ru/article/view?id=39225 - Исаев И.В., Рогачев А.Ф., ФГБОУ ВО «Волгоградский государственный аг-рарный университет»
4. Fundamental research [Electronic resourse] / Интернет-ресурс. Режим доступа: http://www.rema44.ru/resurs/study/sppr/sppr09_expert.pdf - Системы поддержки принятия решений: экспертные методы
5. Сафин Г.Г., Групповая экспертная оценка деятельности операторов одним из методов теории принятия решений - журнал [Текст] / Сафин Г.Г., Абдрахманов А.А., Великанов В.С., Усов И.Г., Савельев В.И., Мацко Е.Ю. «Современные проблемы науки и образования», 2015.
6. Петрянин Д.Л., Структура системы поддержки принятия решений при архивировании данных [Текст]/ Петрянин Д.Л., Юрков Н.К. «Труды Международного симпозиума «Надежность и качество» (2014)»
7. Нзамба Сенуво, Разработка структуры интеллектуальной системы поддержки принятия решений при управлении установкой первичной переработки нефти/ Нзамба Сенуво, Чинакал В.О. «Вестник Российского университета дружбы народов. Серия: Инженерные исследования (2010)»
