Research On Portfolio Investment And Coordination Optimization Of Supply Chain Projects In Logistics Enterprise Based On Multi-Stage Stochastic Programming

Logistics enterprises provide other supply chain members with basic logistics services by their own strength.At the same time,they also gradually involve in multiple supply chain projects investment and operation.However,project operation is often conducted in multiple cycles.What's worse,many uncertain factors exist in the market.As a result,it is the key that how to coordinate multiple supply chain projects for logistics enterprise.Therefore,this paper firstly summarizes the typical supply chain projects in domestic third-party logistics enterprises,extracting the main features of the projects.Secondly,considering the uncertainty of projects income level and the correlation between projects revenue,several multi-stage stochastic programming models are established.Finally,based on the characteristics of multistage stochastic programming models,we use the Cplex,the nested decomposition algorithm and the commercial software--LocalSolver to solve different models respectively,and implement multiple sets numerical examples.The main contents and conclusions of this paper are as follows:(1)We summarize the types of supply chain projects through collecting and sorting the related data in the current domestic third-party logistics enterprises.Then,the characteristics of the uncertainty of projects income level and the correlation between projects revenue are extracted.We also build models including these characteristics.(2)Based on the characteristics of the supply chain projects,this paper establishes multi-stage stochastic programming models.Firstly,considering the case that the degree of coordination between projects is 0-1,an integer nonlinear multi-stage stochastic programming model(INLMSSP)is constructed,and it is transformed into an integer linear multi-stage stochastic programming model(ILMSSP).Furthermore,the coordination degree between projects allowed to be linear rather than 0-1,a bilinear multi-stage stochastic programming model(BLMSSP)is constructed,which is transformed to a linear multi-stage stochastic programming model(LMSSP)by linearization reconstruction.Finally,considering the diminishing marginal utility of project revenue,the Logistic function is introduced as the project's direct income function,and a nonlinear multi-stage stochastic programming model(NLMSSP)is established.(3)During the simulation stage of the model,in the case of the enhancement of the correlation of project income,the correlation between the project income and the total income is always positive correlation in the two models which are the ILMSSP and LMSSP.When the profit space is large or facing the emerging market,the project decision maker should pay attention to the enhancement of the revenue correlation between the projects and the improvement of total revenue.And in the NLMSSP,the correlation between project income and total income is notpositive.At this point,the project decision maker should maintain a reasonable degree of revenue correlation between the projects.In the case of the increase of the correlation among projects revenue and the coordination cost of the projects,in the LMSSP,to balance project benefit and project coordination cost,the degree of coordination between projects decreases with the increase of projects revenue correlation and project coordination cost,resulting in a reduction in total revenue.Therefore,for a project that is sensitive to the change of projects' coordination cost,the decision maker should reasonably control the change of the coordination cost among the projects,and make a plan of the projects' coordination cost through standardized means.When some projects income correlations are enhanced and some projects revenue correlations are weakened,because of the constraint of capital balance and the difference among projects income,not all projects coordinations degree is enhanced.This requires that decision maker should introduce more available funds and coordinate the projects which income correlations are closely related to create more value.After a certain percentage of the initial capital holdings added,as part of the projects return correlations are enhanced and part of the projects return correlations are decreased,in the ILMSSP and LMSSP,due to the increase of initial capital holdings,the amount of funds allocated to each project is changed,so the project capital input and the project income are increased,bringing the increase of total revenue.Therefore,in the face of emerging market or the market profit space is larger,the decision maker can expand the scale of investment and increase project investment to improve total revenue by introducing more available funds.In addition,in the NLMSSP,when the initial capital holdings is increased,the investment and revenue of project have not been improved.At this time,the profit space in the market is limited,and more funds will not bring more benefits.The decision maker should control the amount of available funds within a reasonable range.