The Optimization Of Intelligent Algorithms And Its Application In Logistics Transportation

Logistics transportation is a hot topic today. Along with the diversification of means of transportation, logistics transportation routes choice becomes a NP problem of the combinatorial optimization problems. Whether the logistics transportation routes are reasonable or not directly impact on the speed of delivery of transportation enterprises, the cost of doing business and enterprise efficiency. Researches are made in this thesis on several cases of logistics transportation problem based on intelligence algorithms. The main achievements of this thesis include:1. Firstly, this thesis presents an algorithm of searching for best logistics transportation routes in common instance, which means that the transportation network is a common network. In this part, thesis focuses on how to improve the basic GA to solve the common instance accurately and quickly. The ameliorations on basic GA mainly include: the usage of variable-length chromosomes (routing strings) and their genes (node) for encoding the problem, the setting of alterable crossover and mutation probability, and the in-deeply dividing of mutation operator into gene-insert mutation and gene-delete mutation. Besides, in this part traditional adjacency matrix structure is abandoned, and the linked storage structure is adopted to save space cost.2. In the second part, this thesis puts forward a best logistics transportation route searching algorithm applying to the instance of the network being a common network and existing several source nodes and target nodes.3. As to today's logistics transportation, it does not just constrain to a certain vehicle. The choice of the vehicles is larger and the diversification of vehicles becomes a key point we must pay more attention to when choosing a best logistics transportation route. So the third part of this thesis develops a mathematical model of the distribution of logistics transportation problem based on several model assumptions according to the practice of vehicles.4. The logistic network inside cities is in consideration in this part. Research mainly focuses on lane-network inside cities, which is described as strongly connected directed diagraph in view of its specific characters. The model of strongly connected directed diagraph under certain conditions is optimized by making the digraph d stage. And, d is the period of the strongly connected directed digraph. Based on this model, the algorithm of calculating the optimization route of arbitrary two nodes is designed. Besides, this thesis presents an algorithm of calculating the shortest length of directed H circle. In the end of this part, an example is given to show the feasibility of the two algorithms. This research can be the theoretical base for the logistic optimization of city with lane.5. When the network is small and not so complicate, a lot of algorithms can be used in solving transportation problem. But while network becomes a little more complicate, the asymptotic time complexity of these algorithm will have an exponential growth. So it requires the generation of some algorithms to deal with such complicate network problem. This part indicates some measures which can be described as thought of simplifying of traffic network. There are two exhibitions in this thought. One is to divide the big network into several small networks directly, and dispose of these small networks parallelly, then get the best transportation route by compounding the outcomes of small network. The other is to use the thought in the process of directly solving. In this instance, not all nodes for every step are meaningful, that is to say, in the process of getting the best route, the existence of some nodes are meaningless. So we can make these nodes invisible to simplify calculation and improve efficiency. We use the compound of genetic algorithm and ant colony algorithm to settle this kind of problem.