Research And Application On Vehicle Routing Choice With Ant Colony Algorithm

Modern logistic gradually becomes a new profit source, after natural resource and labor.The most fundamental and important step in logistic is delivery, so how to optimize logistic distribution route is a problem that enterprisers most concerned.lt can radically reduce the cost, so as to reduce logistic costs. Eventually it can reduce the operation cost of enterprises and raise business economic benefits.An improved ant colony algorithm is mainly studied in this paper, and it is used to solve the problem of vehicle routing. The main contents of this paper are as follows:(1) The main contents of the vehicle routing problem and its classification are introduced, the basic model of vehicle routing problems is established, and then some common solutions for this problem are analysed.(2) The ant colony algorithm that solves the vehicle routing problem is introduced, the ant colony algorithm from different aspects is analysed, and several ant colony optimization algorithm is introduced.(3) According to the advantages and weakness of the ant colony algorithm, the ant colony algorithm based on genetic algorithm is introduced.With the analysis and introduction of genetic algorithm, a new algorithm is proposed which is fused by ant colony algorithm and genetic algorithm. Some improvements are proposed as follows:①pheromone updating rule in MMAS is used;②pheromones of the maximum and minimum limited and initialization setted;③state transition rule in ACS is used;④residual factor with pheromone is setted.With these improvements, this two algorithms are integrated well and avoided weakness simultaneously.According to the theoretical basis, vehicle routing generator is developed. Use standard data of Solomon for a new algorithm of testing, in addition with eil51 data to TSP testing.The results show that the improved algorithm has a better performance. Finally the improved algorithm is used in the actual logistics and distribution.According to the actual problem, objective function and constraints are altered, then use the practical data for testing.