Research On Fourth Party Logistics Routing Optimization Problem Under Uncertain Environment

Fourth party logistics(4PL)is an integrator of the supply chain.4PL designs operation solutions for logistics systems to optimize the whole supply chain.With the vigorous development of the modern logistics industry,defects of third party logistics(3PL)about collaboration and integration gradually emerge.Then,4PL attracts wide attention in the industry and academia,and becomes a hot issue in logistics research fields.Routing problem is the key issue of 4PL optimization.Fourth party logistics routing problem(4PLRP)includes two issues:path selection and 3PL providers'selection.Some researchers studied 4PLRP and obtained some results.However,these studies mainly consider deterministic problems.In real-world situation,the logistics operation process could be affected by many factors,such as weather,traffic,man-made wrong operation,and kinds of emergency,and has a strong uncertainty.To make the 4PL system capable in providing efficient solutions to customers under uncertainenvironment,this thesis studies 4PL routing optimization problem under uncertain environment.The contents of this thesis are as follows:(1)4PL routing optimization problem with stochastic transportation time and cost is studied.Influenced by uncertain factors of the objective world,logistics transportation system has strong randomness.If logistics enterprises have historical data to estimate the probability or probability distribution of uncertain events,the logistics transportation time and cost can be described as stochastic variables.Therefore,the 4PL routing optimization problem with stochastic transportation time and cost is studied from the perspective of 4PL.Under the constraint of total transportation cost,an expected value model and a chance constrained programming model are established first to minimize the total transportation time.Then,to improve the effectiveness and robustness of the proposed models,the chance constrained programming model is transformed into an equivalent deterministic model.After that,according to the characteristic that 4PL chooses routes and 3PL providers simultaneously,an ant colony algorithm and an improved ant colony algorithm with replacement strategy are designed to solve the models.Finally,three examples are designed to test the performance of the algorithms and models.Numerical results verify the effectiveness of the improved algorithm,and show that the equivalent deterministic model can guarantee the robustness of solutions as well as the solution efficiency.(2)4PL routing optimization problem considering chance preference is studied.Because of the rapid development of the electronic commerce and the increasing competition in the logistics industry,traditional price competition among enterprises has gradually translated into the competition of the service level.The logistics decisions under uncertain environment always tend to obtain higher customer satisfaction while certain profits are guaranteed.Therefore,the 4PL routing optimization problem considering chance preference is studied from the perspective of 4PL.First,the customer's psychological preference of confidence level about the transportation time and cost is described based on prospect theory,and the mathematic model is established to maximize the prospect value of the confidence levels.After that,an ant colony algorithm and an improved dual population ant colony algorithm are designed to solve the model.Finally,through parameter analysis,the proposed model is compared with the risk neutral model of 4PLRP.Results show that the model considering chance preference can describe customers' psychological preference about transportation time and cost more accurately.Numerical examples also show the effectiveness of the improved algorithm.(3)4PL routing optimization problem considering time preference of customer is studied.Human decisions under uncertain environment are usually affected by psychological factors.To make the 4PL system capable in providing efficient solutions to customers under uncertain environment,achieving higher customer satisfaction under a certain cost constraint,improving enterprise core competition,the 4PL routing optimization problem considering time preference of customer is studied from the perspective of the customer.First,the value function and the weighting function of the 4PLRP's transportation time are described based on the cumulative prospect theory.Then,the prospect value of the total transportation time is formulated,and the cumulative prospect model of 4PLRP is established.After that,according to the characteristic of the problem,an ant colony algorithm and an improved ant colony algorithm are designed to solve the model.Finally,numerical examples show the effectiveness of the algorithm.Additionally,the proposed model is compared with the expected value model and the prospect model considering time preference of customer.Numerical example analysis shows that the cumulative prospect model can describe customers' psychological behavior more effectively and is suitable for groups of customers with different risk attitudes.(4)4PL routing optimization problem with uncertain transportation time is studied.In logistics reality,we sometimes cannot obtain enough historical data to get the probability distribution that events happen.In this condition,due to a lack of historical data of transportation time in 4PLRP,how to describe the transportation time accurately and effectively is significant.The uncertainty theory provides a useful tool to describe uncertain quantities that lack of historical data.Therefore,the uncertainty theory based on belief degree is employed to this thesis,and the transportation time is described as an uncertain variable.The 4PL routing optimization problem with uncertain transportation time is studied from the perspective of the customer.First,under the constraint that the total transportation time's belief degree satisfies the customer's requirement,a uncertain programming model is established based on uncertainty theory to minimize the total cost.Then,the uncertain programming model is compared with the chance constrained programming model which is established based on probability theory to verify the effectiveness of the uncertain programming model.After that,according to the characteristics of the problem,three improved genetic algorithms are designed from the perspectives of repairing infeasible solutions and avoiding infeasible solutions.Finally,numerical examples are designed to illustrate the effectiveness of the improved methods and the applicability of the algorithms.Additionally,compared with the expected value model,numerical examples show the robustness of the uncertain programming model.