EVALUATION OF THE PRACTICAL USE OF SEARCH ALGORITHMS FOR SOLVING LOGISTICS TASKS BASED ON THE TRAVELING SALESMAN'S TASK

Abstract

The relevance of the article is due to the growing practical needs of logistics support in the dynamic environment of service and transportation flows. The need for efficient use of resources during the logistic processing of orders with the use of automation tools is not in doubt today, especially for problems related to combinatorial logic where achieving the optimal result is associated with significant computing resources. One of these tasks is the traveling salesman task (TSP).

The purpose of the publication is a comparative evaluation of a group of algorithms for solving practical logistics tasks, which are based on the task of a traveling salesman.

The research methodology was based on conducting a computational experiment and conducting further analysis of the obtained results for several metaheuristic algorithms, one of which was proposed by the authors as a combination of two well-known methods - the Monte Carlo method and the nearest neighbour method or "greedy algorithm".

As a result of the research, it became clear that the proposed combined method is more effective in solving large-dimensional TSP among the considered algorithms, therefore, its use in practice for solving logistics problems will lead to a more significant saving of resources.

It should also be noted that an approach based on a combination of two, and possibly more, methods can be more productive than using one algorithm, which leaves room for further research.