Research On MDVRPTW Based On Fuzzy Cluster Analysis And Immune Algorithm

The vehicle routing problem has attracted many researchers’ great attentionever since1959, for large scale MDVRPTW problem, to design a quick,effectiveand simple algorithm is of great importance for enterprise to reduce logistics cost,attain the profit and also improve customers satisfaction.Based on this, through reading a large number of documents, summarizes thecharacteristics of MDVRPTW. And through the concrete research on relatedenterprise, introduced the fuzzy clustering thought and immune algorithm to thethesis. Finally used the example to simulate, the result shows that the algorithm isfeasible.The thesis through the analysis of the large scale problem, established the softtime windows mathematical model, proposed three stages as fuzzy clustering,customer dispatch, routing planning to solve the problem.The first phase, using fuzzy clustering method, the basic theory of decisionmaking, combining qualitative and quantitative analysis method, and the use ofmatlab software to divide the customers reasonably; and propose to customersaccording to the20/80rule into key customers, and focus on them, the simulationshowed the algorithm is feasible and effective.The second phase, considering most of the lots are designed to serve ourcustomers, based on this, the thesis used the improved method of nearest distancedistribution to assign the customers to the lot, change MDVRPTW problem intomultiple small VRPTW problems.The third stage, for the problem of small VRPTW, apply designed improvementimmune algorithm combined with other heuristic algorithm to find the path of theschedule, break the traditional experience of the distance from the nearest neighborspractice.When demand change, introduce the idea of interference management to adjustthe path. From the supplier and customer two aspects, established the problem’s interference model, and proceed the simulation experiments.Finally based on Kang Shi Fu drinks’ distribution, analysis the concretesituation， use the designed method to optimize path.