Research On Discrete Glowworm Swarm Membrane Optimization Algorithm To Solve Multi-objective VRPTW

Vehicle Routing Problem(VRP)is a common problem in practical applications such as logistics distribution management and urban traffic improvement.We mainly solve the multi-objective Vehicle Routing Problem with Time Window(VRPTW)in this thesis.At the same time,it minimizes the two goals of the number of transport vehicles and the total distance traveled by the vehicles.By returning a set of non-dominated solutions rather than a single solution,decision makers are provided with more powerful decision support.Multi-objective VRPTW belongs to the NP-Hard combinatorial optimization problem and the existing heuristic methods are difficult to obtain the global optimal solution of the problem.Intelligent optimization algorithms,with their good global search capabilities,are increasingly being applied to solve multi-objective VRPTW.We focus on the discrete glowworm swarm optimization membrane algorithm to solve multi-objective VRPTW in this thesis,and mainly do the following works:(1)Aiming at the multi-objective VRPTW,an NP-hard problem that has important practical significance,and the multi-objective optimization of the glowworm swarm algorithm itself,a modified discrete glowworm swarm optimization algorithm based on time window division(MDGSO-TWD)is proposed to solve multi-objective vehicle routing problems with time windows(MVRPTW).In MDGSO-TWD,customers are divided into different classes according to their time windows,and the way of generating individual glowworm swarm is thus designed.In order to more realistically and effectively reflect the evolutionary mechanism of the glowworm swarm algorithm,we redefine and design the movement rules of the it,and use random individuals to replace the repeated individuals in the population during the search process to maintain the diversity of the population.At the same time,in order to speed up the search speed and make up for the lack of local search ability of the algorithm,a meaging neighborhood search technology based on time window classification and adaptive glowworm swarm flight step length are proposed to enhance the local search ability of the algorithm and during the search process,taking an elite strategy to record the global optimal solution to prevent algorithm degradation.Due to the non-evaluability of multi-objective problems,the concept of Pareto dominance is introduced to evaluate the pros and cons of individual glowworm swarms,and the way of generating Pareto non-dominated solutions is redesigned.Finally,the effectiveness and efficiency of the proposed algorithm are verified by simulation experiments and comparative experiments.(2)On the basis of improving the discrete glowworm algorithm,by analyzing the three elements of the membrane computing,model's object,reaction rules and membrane structure,the glowworm swarm optimization algorithm based on membrane computing is established(PDVA),which not only retained the basic of the general membrane computing system characteristics,also absorbed the research results of the glowworm swarm algorithm.The algorithm is based on a degree 3 tissue P system in which glowworm swarm coding is used as the evolutionary object,and discrete glowworm swarm evolution mechanisms and variable neighborhood evolution mechanisms are used as sub-algorithms to evolve the objects.In order to enlarge the search space of the solution space and improve the accuracy of the solution,the evolutionary mechanism of variable neighborhoods is improved.Finally,the results of different Solomon calculations show that the PDVA algorithm can be compared with other intelligent optimization methods.