Research On Algorithm For Solving Vehicle Routing Problem With Hard Time Window

With the rapid development of modern economy,especially the rapid rise of e-commerce platform,make logistics distribution an indispensable part of it,and people's demand for logistics distribution services is increasing.As an important issue in logistics distribution industry,vehicle routing problem has attracted extensive research from operational research and portfolio optimization.With the increasingly stringent and timely delivery requirements of customers,the vehicle routing problem with hard time windows is becoming more and more prominent,and the research of this problem has not been well solved due to its complexity.Therefore,this paper carries out research based on this,and the specific research contents are as follows:1.An improved fireworks algorithm is proposed for solving vehicle routing problems with hard time windows.The algorithm can use information interaction for resource allocation.Firstly,based on the traditional fireworks algorithm,the initial fireworks and path solution are generated by combining the construction algorithm,and then the explosion operator of the traditional fireworks algorithm is improved,and then the explosion operator is improved,so that the fitness value of the fireworks population can be adjusted adaptively according to the individual fitness value.Ability,then use cross-reorganization to complete the neighborhood search of the explosion spark,and enhance the diversity of the population through the mutation operation.Finally,the distributed information sharing mechanism of the fireworks algorithm is used to avoid the premature algorithm.2.Aiming at the problem that the improved fireworks algorithm has a long total driving distance when solving the customer cluster test set,a solution method based on the meme algorithm is proposed.Considering the convergence of the initial solution affecting the convergence speed of the algorithm,the initial feasible solution is first generated by fuzzy clustering to ensure the diversity of the population and the convergence of the algorithm.Then the evolutionary operator of the evolutionary module is improved.The edge crossover recombination operator is used for global search to generate more optimal individuals.Finally,in the local optimization module,node exchange and k-opt are used as meme operators to optimize the better individuals to converge to the optimal solution quickly.3.The Solomon standard test set is used as the current international reference standard for vehicle routing problems,which facilitates the centralized comparison ofvarious algorithms.Based on the test set,this paper carries out relevant experimental research on the above main research methods,and compares the results with the published optimal solutions to verify the effectiveness and feasibility of the proposed algorithm.4.Finally,taking the SF Express Lanzhou Distribution Center as the example for the delivery of express delivery,this paper is based on the simulation experiment of Flexsim simulation software.The experimental results show that the proposed algorithm has certain theoretical significance and application value for solving vehicle routing problem with hard time windows.