Research On Multi-Chamber Vehicle Scheduling Optimization Based On Improved Genetic Algorithm

With the further research on vehicle routing problems in recent years,the multi compartment vehicle routing problem has become an important research direction.In real production and life,scenarios such as finished oil distribution,garbage recycling,and cold chain transportation all require the use of multi compartment vehicles for transportation or distribution.The openness of China’s refined oil market is deepening,and the competition in the oil market is becoming increasingly fierce.In the product oil sales enterprises,the crude oil produced from the oil field is transported to the refinery through the Oil pipeline,which is called upstream transportation;The process that refined oil is transported in refining and chemical enterprises and delivered to the oil depot of refined oil sales enterprises through railway,waterway or Pipeline transport,and then unloaded and warehoused is called one-time distribution;The process of transporting oil products from the oil depot of the product oil sales enterprise to the gas station or customers by road through Tank truck is called secondary distribution.The proportion of the total logistics cost of the road distribution cost of refined oil products is as high as 60-70%,and in most cases,the distribution mode of refined oil products mainly depends on multi compartment Tank truck to transport on land.The controllability of refined oil secondary distribution is relatively high,and the logistics costs it bears are one of the main sources of logistics costs for petroleum enterprises.Therefore,refined oil secondary distribution has become a major breakthrough for petroleum enterprises to reduce costs.Firstly,this article takes the problems that occurred during the secondary distribution process of a certain oil distribution center as the background,and combines the actual needs of a certain oil distribution center.The optimization objectives are to minimize the number of distribution vehicles,minimize vehicle transportation costs,and minimize the time penalty cost for vehicles violating customer time windows.A multi-objective comprehensive optimization function for the MCVRPTW problem is constructed,and a mathematical model is established.Secondly,based on the standard genetic algorithm,this article further optimizes the MCVRPTW problem and proposes an improved genetic algorithm for solving the MCVRPTW problem.In the improved genetic algorithm,the nearest neighbor matrix and Adaptive Neighborhood Method(ANM)are first used for population initialization.The improved roulette wheel selection method is used to improve the superiority of the initial population.The adaptive greedy crossover method and parallel crossover method are introduced into the crossover operator,and the fierce wolf attack operation of the wolf swarm algorithm is also introduced to provide direction for the crossover operation of the genetic algorithm,Improve the optimization efficiency of the algorithm,introduce inversion operations and variable neighborhood search algorithms into mutation operators,reduce the uncertainty of genetic algorithm mutation operations,and further expand the search range of mutation operations to improve the convergence of the algorithm.Finally,combined with a population update strategy based on group similarity,the population that has completed the operation is further updated.The improved genetic algorithm was applied to the practical problem of secondary distribution of refined oil products,and the good performance of the improved genetic algorithm was verified.