Research on the Hybrid Fruit Fly Optimization Algorithm with Local Search for Multi-compartment Vehicle Routing Problem

With the rapid development of modern logistics industry, multi-compartment vehicle routing problem (MCVRP) is one of the research directions of vehicle routing problem, mainly to solve the problem of the transportation of non-mixed products in one vehicle. These products must be stored in separate compartments, and should not exceed the capacity of the carriage. A hybrid fruit fly optimization algorithm (HFOA) is proposed to solve the multi-compartment vehicle routing problem with capacity constraints, which combines three local search methods: 2-OPT, Swap and Insert. In HFOA, firstly, the initial feasible solution of MCVRP is constructed by using the stochastic method. Secondly, from the initial solution, the path probability function is used to create the fruit fly path. Then, three local search methods are used to optimize the path, and the optimal solution is used to update the trajectory strength of the distribution network. To improve the search efficiency, the local search process is only implemented in one iteration. Based on the simulation results of 7 benchmark problem, it is found that the HFOA algorithm can produce better path planning than traditional methods.