A Multi-Step Algorithm for Finding the Optimal Strategy for the Behavior of a Group of Robots Based on Intelligent Methods

摘要

The problem of multi-criteria search of the optimal plan of tasks for a group of mobile robots is considered taking into account the energy and time spent by each of robots, as well as the number of tasks performed. Earlier authors described the applicability of genetic algorithms (GA) to the simpler problems of this class. This article proposes to use the GA to optimize a task with several criteria and the ability to find an optimal strategy for the behavior of a group of robots with several steps ahead. The trajectories of group's robots are planned using a neural network algorithm providing neural maps. The method of synthesis and realization of fitness function taking into account the specifics of the problems to be solved and the type of GA is described in detail. The results of the simulation are presented, in the course of which strategies for the behavior of a group of robots with different source data are generated. The advantages of this approach are shown in comparison with the previously implemented one-step optimization algorithm.