Application Research On Vehicle Routing Optimization Problem Based On Genetic Algorithms

Vehicle routing optimization problem has been proved to be a NP-Hard problem with multiple constraints.It is also one of the hot spots in various countries,Because of its constraint condition,it is difficult to solve the problem.On the basis of reading a lot of literature,it is found that there are many problems in the process of logistics distribution,and the single objective distribution is more,but in real life,there are different requirements for different customers.The single goal was too vague and thus ignore the customer's individual needs,so this paper takes into account the needs of different customers,set up a wide range of goals on transport costs,the number of transport vehicles,customer satisfaction,combined with the distribution process to avoid vehicle circuitous transport,customer satisfaction within the specified time window can receive the actual requirements of the goods,the establishment of a static multi-objective mathematical programming model that can better describe modern logistics and distribution problems.This thesis is from the perspective of multi objective vehicle routing optimization,we summarize the related theory and basic model of the vehicle routing optimization problem by using the literature retrieval method,combined with the actual situation of Tianjin Seven enterprise distribution,establish a multi-objective model with the actual distribution needs.Through the comparative analysis method,the method is used to solve the multi-objective model,which is suitable to solve the multi-objective model of the fast non dominated sorting method,namely NSGA-? algorithm for the final path of the solution.However,NSGA-? boils down to is a genetic algorithm,it is impossible to completely avoid premature convergence of genetic algorithms as well as the strong dependence of the shortcomings of the initial population,So this paper use the classical path optimization method Or-opt combined with NSGA-? algorithm,the method of path optimization for multi-objective vehicles is improved.In the process of solving the model,the relevant reliable distribution basis,including customer's geographic coordinates,demand,delivery time window,the unit cost of loading and unloading,as well as the unloading time,as well as the penalty costs for the distribution enterprises are collected.Finally,MATLAB simulation is used to verify the feasibility of the improved NSGA-? algorithm,and the convergence rate of improved NSGA-? is faster and the convergence curve is more stable than that of NSGA-?.Finally,a reasonable Pareto optimal solution is obtained by the combination of enterprise distribution,which meets the requirements of vehicle routing optimization.