Development of algorithms for the information system for constructing the optimal delivery route

Brykin Dmitry Olegovich, Pestova Svetlana Yurevna

The article presents the results of the study of the speed and optimality of the algorithms for solving the traveling salesman problem for companies that need to optimize their own delivery service. Five algorithms were analyzed and three of them were compared. For the subsequent work, it was decided to use an exhaustive search algorithm for short routes and a combination of simulated annealing and genetic algorithms for longer routes. Algorithm tests have shown their strengths and weaknesses. The algorithms were evaluated for their speed and the resulting route length. The genetic algorithm and the simulated annealing algorithm have shown good efficiency if the number of points in the route is large. For short routes, a brute-force algorithm is better suited. It was decided to use their combinations depending on the calculated route.

Keywords:logistics, routing, optimization, traveling salesman problem, metaheuristic algorithms.