Study On The Structure Complexity And Design Optimization Of Logistics Networks

Logistics network, as a carrier of logistics activities, is the physical foundation of the logistics managemnt. Logistics network is also a complex system, in which many optimized problems need to be solved. Hence, to study the logistcs network with complex system and optimization theories could understand runing rules of logistics activities further well and provide a new effective method and scientific foundation to research and plan the network. This dissertation mainly disscusses the complexity of logistics networks and analyses its price evolving characteristics. Logistics network design models with different conditions are proposed and logistics vehicle routing problem with fuzzy constraints is studied.The main content and innovation of this dissertation are summarized as follows:(1) Based on complex network theory, the bearing capacity of logistics networks with different topologies is studied, and optimal network structure is obtained. By considering the elastic demand condition, we analyze load distributions and bearing capacities with different parameters through artifcially created scale-free networks. The simulation results show that the load distribution follows a power-law form, which means some ordered pairs, playing the dominant role in the transportation network, have higher demand than other pairs. We also have found that, with the decrease of perceptual error, the total and average ordered pair demand will decrease and then stay in a steady state. However, with the increase of the network size, the average demand of each ordered pair will decrease, which is particularly interesting for the network design problem.(2) The spatial price problem means that if the supply price plus the transportation cost is less than the demand price, there exists a trade. And vice versa. This paper studies the evolution results of trade price in different network structures. Simulation results show that the greater the degree of random network connection and the more obvious the community structure are, the average price down faster, the system will reach equilibrium in a short time. For a scale-free network, the larger of the network size, the slower of the average price falls. These results show that the network with shorter path length is sensitive to the variation of prices.Taking into consideration the characteristics of their competition of the logistics service provider and customers, game theory and bilevel programming are used to describe the pricing of logistics services. And application example of the model and algorithm are tested, the results show that the algorithm has good convergence, other relevant parameters have different effects on the upper and lower objective function values.(3) By taking the flow uniformity into account, the location model is established to reduce the logistics resource waste and to distribute the flow uniformly. A simple solution algorithm is given and the example shows that the model is effective. Sharing market information has been recognized as an effective approach to reduce demand distortion and improve supply chain performance. This paper proposes a bilevel model to solve the location problem by considering information sharing effects. The results show that with the increase of the information sharing, the total system cost will significantly reduce. But with the further increase of the level of the information sharing, the system cost will be gradually stable. This shows that the level of information sharing will affect the interests of the logistics business and customers, a reasonable degree of the information sharing can also improve the profits of both.(4) Logistics vehicle routing problem with fuzzy constraints is studied. On the basis of previous studies,the optimizing model of logistics vehicle routing problem with fuzzy chance constraints is established, and a fuzzy analogy genetic algorithm for solving this problem is proposed. Finally, a simple experimental example is given. The result shows that this algorithm is a valid method to solve the disdeterminate problem, and have good search optimal capacity.