Application of Hybrid Fruit Fly Optimization Algorithm in Multi-robot Path Planning

Robot path planning is one of the key technologies in the field of autonomous robots, and the path planning problems of multiple robots are even more challenging and far from being fully researched. We propose a Hybrid Fruit Fly Optimization Algorithm (HFOA) based on the original Fruit Fly Optimization Algorithm to solve these problems, considering to overcome its shortcomings such as being prone to local optima and lack of balance between global and local search. Firstly, a Halton sequence-based initialization strategy is adopted to obtain a higher-quality initial population. Secondly, to increase the diversity of the population, parallel search and sine-cosine search strategies are introduced. The whole population is divided into two subpopulations, which evolve by different position update strategies and exchange information periodically to effectively balance the performance of global and local search. Finally, the Metropolis acceptance criterion is adopted for individual update, enhancing the ability of the algorithm to escape local optima. Based on path planning of two robots on two typical maps, the mathematical optimization model for the path planning is established and the following problem-solving strategy is proposed. First, Dijkstra algorithm is applied to obtain the coarse-grained planning, then the proposed HFOA is employed to solve above optimization model and obtain the final planned paths. Through comparison with the results of related algorithms and data analysis, the feasibility and effectiveness of the proposed algorithm in solving such multi-robot path planning problems are verified.