The Traveling Salesman Problem Of Logistics And The Study Of Heuristic Algorithm In Agricultural Products Distribution

At present, the agricultural products logistics transport is used widely, but theoretical research, especially, systematic research is few. Vigorously developing agricultural products logistics has extremely strong realistic meaning, it can decrease circulation cost of agricultural products and improve the value of agricultural products. This paper proposes a traveling salesman problem of logistics constraints on the issue of the agricultural products logistics distribution. In this study car starts from product base, going through all the retail locations and meeting the demands of agricultural products for retail locations, finally it returns to vegetable base. We are trying to find a closed path that makes the range of product of vehicle load and distances minimum.Agricultural products logistics distribution problem is regarded as a variation of traveling salesman problem(TSP). Traveling salesman problem is a classic NP complete problems, it has wide and practical application value. Solving TSP problem with classical methods mainly includes: greed, branch and bound method, dynamic programming, nearest neighbor heuristic, Hopfield neural network optimization algorithm, ant colony algorithm, simulated annealing algorithm, genetic algorithm and hybrid optimization strategy. Based on the connection between logistics distribution problems and TSP problem, the genetic algorithm was improved, we figure out an initial solution using the methods of TSP chain on the basis of the known optimized solution, thus we can get a desirable solution. We consider the demand of R influence on the optimal solution and try to make R multiplied by a factor of demand. So we can make it increasing at 0- R, at the same time of increasing. We use each optimal solution that is computed as the initial solution. On the basis of the optimal solution, we can make the algorithm optimize again. At last, we find the optimal solution. Through experimental analysis, the results show that the algorithm is effective and can obtain good results, it also improves the efficiency of the algorithm.