Research On Invulnerability And Network Optimization Strategy Of Urban Logistics Network

Recently, urban logistics network played a more important role in ensuring the normal running of urban economy and urban people’s livelihood with the development of social economy and urbanization in China. But emergent events such as 7-10’ heavy rain in Beijing, 2010 Shanghai Expo, holiday promotions of e-commerce, rural workers’ rushing home challenged the stable running of the urban logistics network by increasing in logistics demand and reducing of logistics supply suddenly. Then, whether urban logistics network can ensure the necessary functions of keeping the normal running of urban economy and urban people’s livelihood after emergent events has been one of the important factors for the planning and construction of urban logistics network, promoting the modernization level of urban logistics and urban living standard. The following four aspects have been carried out for the purpose of enriching urban logistics network planning theory, supporting the urbanization in China, constructing the method and model, exploring the optimization method of urban logistics network’s invulnerability.(1) Research on the evolution mechanism, model and algorithm of urban logistics network.The evolutionary process of urban logistics network was researched. A weighted BA model was proposed based on abstracting the economic evolution mechanism as minimizing the cost of the network. After designing the evolutionary algorithm and simulating of the process of agglomeration and diffusion evolution of urban logistics network, the basic characteristics of evolution were discussed, providing a base model for invulnerability analysis.(2) Research on the static invulnerability of urban logistics networkBased on the basic model (1), a static network model composed of nodes, edges and logistics volume was constructed, ignoring cascading failure temporarily. Macroscopically, considering the integrity of the area covered by network, microscopically, considering the service maintenance of nodes, a system of measuring was designed including the size of largest connected subgraph and average clustering coefficient. Selecting node failure as the failure type, random and selective failure as the failure mode, point strength, cross entropy, improved betweenness and random as the failure strategy, the simulation process was planned and the simulation results were analyzed from the perspective of different failure strategies and different network sizes. Finally, the corresponding network optimization strategies were proposed based on average degree and strength of nodes, evolutionary distance weight of nodes.(3) Research on the dynamic invulnerability of urban logistics networkBased on the basic model (1) and (2), a dynamic network model composed of nodes, edges and logistics volume was constructed based on cascading failure, including the mechanism of capacity allocation, node status, load distribution and failure propagation. Considering to measure the damage, cost and influence, a system of measuring was designed with node survival rats, network cost effectiveness and cascading failure number. Selecting node failure as the failure type, selective failure as the failure mode, improved betweenness as the failure strategy, the simulation process was planned and the simulation results were analyzed from the perspective of different capacity reserve coefficient and overload failure probability. Finally, introducing different capacity allocation mechanisms and different load redistribution mechanism in network optimization, the corresponding network optimization strategies were proposed based on prevention before failure and emergency response after failure.(4) Empirical study:a case study of Chengdu CityConstructed Chengdu city logistics network of 413 nodes,1456 sides by calling Baidu Maps API. Static invulnerability was analyzed through different failure strategies and evolutionary distance weight of nodes. Dynamic invulnerability was analyzed through different capacity reserve coefficient, failure probability, capacity allocation mechanism and load distribution mechanism. Finally, network optimization strategies were proposed based on prevention before failure and emergency response after failure.At last, the research prospects were proposed to two aspects, the introduction to geographical location parameters in evolution models of urban logistics network and the combinatorial optimization of reserve capacity, capacity allocation and load redistribution.