Two-stage Optimization For Integrated Production-distribution Planning Under Vehicle Operational Restrictions

In make-to-order supply chains,the end-customer oriented Distribution Center in charging of order assetnbly(production)and distribution is one of the core departments,whose operating performance would largely affect the supply chain profit and customer service level.Especially,this production and distribution operations are high related to each other,thus its production-distribution planning in such supply chains is one of the key problems in operations management.However,in practice,this system is often affected the uncertainties,including the fluctuation of customer demand,delivery due time,order delivery time et al.Furthermore,in many Chinese metropolitan areas,in addition to normal demand and traffic uncertainties,the operability of each vehicle in the fleet can be uncertain due to randomly occurring severe weather condition(e.g.,smog),which leads to imposition of vehicle operational restrictions.These uncertainties make the daily production and distribution resource availability and demand become uncertain.Therefore,to achieve the production and distribution coordination under these uncertainties,we should not only improve the production and distribution scheduling decisions in the operational level,but also optimize the production equipment and distribution resource configurations in the strategic level.Since the integrated production-distribution planning under vehicle operational restrictions contains these two coupling decisions,it is featured as a dynamic,unstructured and complex problem.This research proposes a 2-stage optimizing method for integrated production-distribution planning problem based on sequential decision-making methodology by decomposing it into resource planning and operation scheduling problems.Furthermore,it integrates these two problems by a 2-stage stochastic integer programming method for the strategic problem and an integrated scheduling method with given strategic planning for the operational problem.This work can help achieve effective plans for supporting the decision-making process of production and distribution coordination.The research includes the following aspects:(1)Problem statement and decision-making analysis for production-distribution planning problem under vehicle operational restrictions.First,this section describes the production-distribution planning problem under vehicle operating restrictions.Then,it analyzes the demand and delivery uncertainties in daily operations and their affections on corresponding decisions.Finally,it analyzes the decision-making process of production-distribution planning problem and proposes the 2-stage optimizing method for it.(2)A 2-stage stochastic programming for integrated production-distribution resource planning under vehicle operational restrictions.First,it generates the delivery scenarios and corresponding stochastic parameters via the Monte Carlo sampling method.Then,it formulates a stochastic programming model that aims to minimize the strategic-level procurement spending on production equipment and delivery fleet and the expected operational cost under logistic uncertainty.In order to solve this problem,the research adapts a stochastic branch-and-bound(SBB)algorithm,for which a local search heuristic is applied to improve the efficiency of the upper bounding and a scenario grouping method is applied to reduce the computing burden.(3)Integrated production-distribution scheduling under vehicle operational restrictions.First,this research extends the production routing problem to make-pack-route problem to devise production and routing schedules that allow flexible order composition and ensure in-time delivery.Then,it develops an iterative sequential scheduling heuristics algorithm embedded with local search,which solves the three corresponding scheduling problems sequentially and improves the overall schedule in an iterative manner.Especially,to reduce the computing burden,two grouping and sequencing heuristic algorithms are developed for routing and making problems,respectively.(4)Numerical experiment and case study.This section presents the numerical experiments results based on the real case of a fresh produce online retailer in Beijing city and the real-world traffic data collecting from Baidu map.It verifies the proposed decision-making methods and evaluates the performances of the proposed algorithms.It also offers managerial insights derived from these computational results.This study integrates the methodologies of sequential decision-making,stochastic programming and integrated scheduling and makes a beneficial exploration for the production-distribution planning problem under vehicle operational restrictions.It offers a new methodology for describing and solving the similar dynamic,unstructured and complex problems containing strategic and operational level decisions under uncertainty.It also extends the current research boundaries of stochastic programming theory and integrated scheduling methodologies.This study would also help solve the production and distribution coordination problem under uncertainty and offers a solution for city delivery problems caused by the vehicle operational restrictions due to smog,which is significant to the survival and development of related industries.