The 4PL Network Optimization Problem Considering The Risk Of Demands

Logistics is the process which goods are sent from supplier to receiver along with transportation, storage, loading and unloading, handling, packaging, information processing. Along with the advancement of economic globalization and the social enterprises’requirement of logistics services is increasing, the traditional third-party logistics cannot meet the requirements of the enterprise, therefore, the fourth party logistics arises at the historic moment.The fourth party logistics is the provider who provides a complete set of supply chain solutions to its clients, it integrates the different resources, capabilities and technology of clients and third party logistics, so as to improve the efficiency of the supply chain. The fourth party logistics focus on all aspects of the supply chain, it breaks through the limitations of the third party logistics and make information sharing a reality, thus can make the operation of high efficiency and low cost of supply chain.In this paper, we study the 4PL network optimization problem considering the risk of demands, analysis the influence of the random demand situation for the result of the experiment. According to the attribute, constraint and research target of fourth party logistics, we use the thought of multi-graph to build a mathematic model of the fourth party logistics problem. When we know the demand’s distribution function, we can transfer the uncertainty model to a certainty model, then we can use CPLEX to solve the model;when we get a little knowledge about the demand’s distribution function, we can solve the uncertainty model through Improved Particle Swarm Optimization Algorithm by using the history data。We verify the suitability of improved particle swarm optimization algorithm towards different sizes of problem. Through the comparison between improved particle swarm optimization algorithm and traditional particle swarm optimization algorithm, we verify that the improved particle swarm optimization algorithm has obvious superiority. Then, we realize the simulation experiment of different sizes of the fourth party logistics’optimization problem, We contrast the results of the models of the demands’ distribution functions are known and unknow, and make comparison to verify the influence of the mutation rate of demand distribution and the service level of client’s requirement about the fourth party logistics’optimization problem. We also consider the influence of the number of history data on the improved-pso’s results.