Research On Reliable Location Problems For Regional Logistic Network With Information Failure

Traditional facility location problems try to find a set of reasonable facility locations in a certain service area in order to satisfy customers’ logistic demands with the optimal system cost.With the development of research,people realize the importance of system reliability.The reliable location models with the probability of node failure become the hot topic in the location research.In the same service area,the failure or hysteresis of information network often happen with node failure and influence the choice of customers’ path especially in regional logistic network with high correlation.According to the characteristics of regional logistics network with high correlation,considering the node failure and diffusion phenomenon which have become the main constrain factors for the reliability and flexibility of logistics network,this paper proposes a reliable location optimization model under information failure scenarios which is conforms to the characteristic of logistics network.Design a reasonable algorithm and discretization algorithm.We analyze the validity and inner characteristics of the proposed model by numerical examples.Then we make a analysis for Jing-Jin-Ji region which is the typical regional economic in our country,reveal the practicability of the proposed model,analysis the main factors influencing the regional logistics network performance,and explore the operation methods and management strategy for controlling control the reliability and flexibility of regional logistics network.These studies not only enrich the theoretical results of the node location research of logistics network nodes,but also provide a theoretical method for the node location planning.The main research contents of this paper are shown as follows:(1)Continuous reliable location model under imperfect informationWith the assumption of real information failure and considering the particularity of logistic activity,two-way trip,we propose a continuous node location model.Firstly,we describe the continuous node location problem with imperfect information and two-way trip.Besides the normal variable,node built number,we illustrate two particular decision variables:node built location and the customers’visiting sequence for service.We propose our first innovation,"trial and error" strategy.In this strategy,customers know the locations of nodes but have no information about the real status of nodes.Therefore,customers must visit the pre-assigned nodes until they find the first operational facility for service or give up looking for the service,and then they return to their initial locations.Secondly,according to the description of CRLP-ⅡTT,the section defines the cost components of total system cost from the view of mathematics including node construction cost,customers’ transportation cost and penalty cost,and then proposes CRLP-ⅡTT model.The aim of the proposed model is to obtain the optimal total system cost by selecting the suitable node location scheme and customer’s visiting sequence.Because CRLP-ⅡTT model is hard to be solved,CRLP-ⅡTT model in the heterogeneous plane is transformed by a continuous approximation method.The decision variables is also changed to the initial service area of node from the suitable node location scheme and customer’s visiting sequence.The near-optimal solutions of CRLP-ⅡTT model is obtained by solving the transformation model.Finally,this section compares the optimal solutions with perfect and imperfect information,two-way and one-way trip respectively.The study results indicate that it is necessary to consider the imperfect information and two-way trip.The sensitivity analysis of key parameters is carried out to validate the robustness of the proposed model.(2)Discretization algorithm design for continuous modelThe near-optimal solution,which is obtained by the continuum approximation method for the proposed continuous model,is the node initial service area and built number.This method is unable to obtain node location scheme and customer visiting sequence directly in the planning area.Therefore,the node location scheme should be obtained by the discretization algorithm based on the node initial service area.And then the customer visiting sequence is obtained.According to characteristics of CRLP-ⅡTT model,a customized discretization algorithm is developed based on cellular automaton and the corresponding code is compiled by Matlab software.The efficiency of customized algorithm is verified by a numerical example.(3)Optimization the reliable node location for the logistic in Jing-Jin-Ji regionIn order to verify the ability of CRLP-ⅡTT model for solving the realistic problems,this paper optimizes the logistic network nodes in Jing-Jin-Ji region.Firstly,we make a survey and obtain the discrete data from Jing-Jin-Ji regional logistic network.Based on the discrete data,we get the continuous data to formulate the parameters of continuous model.Secondly,solve the CRLP-ⅡTT model for Jing-Jin-Ji regional logistics nodes to obtain the optimal initial service area for any node in this region.Then the customized discretization algorithm is used to obtain the node location scheme.Finally,adjust the partial node location scheme to obtain the practical logistics node location scheme in Jing-Jin-Ji region.On the basis,analyze the influence of various parameters in the proposed model on node location scheme that obtains some beneficial thought for node location in Jing-Jin-Ji region and provides certain references for the coordinated development of Jing-Jin-Ji region.(4)Discrete reliable location model under budget constrainThis chapter proposes a discrete reliable node location model under imperfect information and budget constrain.A customized Lagrangian relaxation algorithm is developed for solving the discrete proposed model.The Jing-Jin-Ji regional logistics node data is used for example analysis.Research results show that the proposed model and customized algorithm can solve the logistics node location problem in Jing-Jin-Ji region to obtain reasonable node location scheme.Through the sensitivity analysis of model parameters,study the influence of various parameters on the optimal total system cost and verify the robustness of the proposed model.