Task Allocation Method of Multi-Logistics Robots Based on Autoencoder-Embedded Genetic Algorithm

This work considers a task allocation problem of multi-logistics robots in closed campus environments. To find its high-quality solution in a reasonable time, an autoencoder-embedded genetic algorithm is proposed. In it, genetic algorithm is selected as the basic solution framework. In order to deal with the difficulty of constructing 'building blocks' in a high-dimensional solution space, an autoencoder network is introduced to compress a high-dimensional solution into a low-dimensional one. Then, genetic algorithm can construct its 'building blocks' in the resulting informative and low-dimensional solution space. Hence, a parallel framework involving two co-evaluated subpopulations is constructed. Genetic algorithm works in both the original solution space and the low-dimensional one generated by the autoencoder network. Simulation results based on some instances and comparisons with some existing algorithms demonstrate the effectiveness and robustness of the proposed algorithm.