Research On Fourth Party Logistics Routing Problem Considering Earliness/Tardiness Penalty With Stochastic Time

With the development of Economic Globalization,the company's demand for logistics continue to increase,as a result,the Fourth Party Logistics(4PL)was born on the basis of traditional logistics.The Fourth Party Logistics,as a supply chain integrator,provides logistics services to customers by selecting the Third Party Logistics(3PL)providers and transit nodes.In order to reduce logistics expenses,the first problem is to optimize the distribution route.In previous studies,the logistics network was abstracted into a multi-graph.Each arc between two points corresponds to a 3PL provider,and each intermediate node represents a transit city.The Fourth Party Logistics Routing Problem(4PLRP)is to find a toute scheme which can meet the constraint conditions and make the objective function get the most value from the supplier point to demand point.Most of current research is standing on deterministic problems without considering influences by practical changes.Based on the the Fourth Party Logistics Routing Problem,stochastic planning and ant colony algorithm,the paper studies the randomness of the transportation time due to various uncertainties in the logistics transportation and the penalties caused by the logistics service being unable to be completed on time due to the randomness of transportation time.The specific research content as follows:(1)The paper studies single point to single point 4PLRP with single task.The logistics transportation time has a strong randomness influenced by the uncertain factors of the objective world.Contrapose the randomness of transportation time,the Expected Value Model(EVM)and the Chance-Constrained Programming Model(CCPM)are respectively built to minimize the cost of logistics transportation.For 7 nodes problem,use CPLEX and MAX-MIN Ant System(MMAS)solve the Expected Value Model respectively,and the results of MMAS were compared with those of the CPLEX solution to verify the effectiveness of the intelligent algorithm.Combine stochastic simulation with MMAS to solve the CCPM.And use the intelligent algorithm to solve 15 nodes and 30 nodes problem.Compare the results of two models.The conclusion shows that the result of the Chance-Constrained Programming Mode is better than the Expected Value Model.And the result of CCPM can be adjusted dynamically according to the different requirements of the client,but the computation time is longer.At the same time,the influence of different factors on the solution of the two models was discussed in the study.(2)The paper studies single point to single point 4PL optimization problem with multi-tasks.Consider that 4PL service providers undertake multiple logistics delivery tasks at the same time,and each 3PL service provider can only serve one transportation task and build the Chance-Constrained Programming Model.According to the characteristics of multi-task problems,improve ant colony algorithm to solve the model.The result shows that the improved algorithm can reduce the relative standard deviation of results and improve the stability of the calculation results after comparig the improved algorithm with MMAS.