Research on Multi-Robot Task Allocation Based on BP Neural Network Optimized by Genetic Algorithm

Abstract

Performing time and speed in multiple robots search and rescue is the key, requiring multiple robots to quickly complete task assignments. When each robot bid for a task with multiple prices, the BP neural network algorithm is used to fuse these bid prices and determine the robot that wins the bid. However, the traditional BP neural network has the disadvantages that network training can easily fall into local minimums and training errors. It will lead to long time and slow speed for multiple robots to perform tasks. Based on the traditional BP neural network algorithm, this paper proposes a method based on genetic algorithm to optimize BP neural network. This method can effectively deal with the problem of redundant parameters, and the training error is small. The simulation results show that the proposed method can greatly improve the training times and errors of BP network. It also verified that multiple robots can save time when completing tasks.