METHOD OF GENETIC PROGRAMMING FOR SOLVING THE PROBLEM OF OPERATIONAL SCHEDULE PLANNING OF DISCRETE PRODUCTION

Abstract

One of the main conditions for the successful functioning of the enterprise is a well-organized production planning process. Production planning systems of the APS/MES class, the basis of which are algorithms for building production plans, allow automating this activity. The paper examines the problem of scheduling for enterprises of a discrete type of production, related to the field of multi-criteria optimization problems. A formal description of the planning task is given, taking into account the main production constraints (time constraints, equipment requirements and the order of operations). The main methods of solving problems of this class are briefly considered; their main advantages and disadvantages are noted. To solve this problem, an approach based on the generation of heuristic rules used in planning production operations for specified resources has been chosen. Based on this approach, a two-stage algorithm for building production schedules is proposed, which includes the generation of dispatching rules and their further application in building schedules. A genetic algorithm is responsible for generating dispatch rules. The implementation of its genetic operators is described in detail, as well as the composition of the chromosome and the tree representation of the dispatch rules included in the chromosome. The algorithm is implemented in C# 12 using a free platform.NET 8. The implemented algorithm has shown its effectiveness in comparison with the greedy algorithm on small generated datasets. Further research in this area is aimed at evaluating the effectiveness of the constructed algorithm with more complex genetic operators and the structure of the expression tree, as well as reducing the duration of the process of generating heuristic rules for large data sets.