Optimization Of Perishable Goods Reverse Logistics, Inventory Control And Vehicle Routing Problem

Supply chain management(SCM) has received tremendous attention from the business world and form academic researchers in the past 20 years.But,most of the SCM researches concentrate on the forward movement and transformation of the materials from the suppliers to the end consumer.However,the reverse flow of products from customers to upstream businesses has not received so much interest.The importance of studying reverse logistics has increased in recent years for many reasons,such as the deteriorating environment and the lack of the natural resources,more and more the relative laws that force our company be responsible for the whole life time cycle of their products,customer expectations that urge companies to reduce the environmental burden of their product,and the economical feasibility due to material and added value recovery.Reverse logistics has become a key competence in modern supply chains.Issues arising encompass distribution,inventory and product management aspects.The previous researches about reverse logistics can be classified into four categories:(1) inventory management in the system with return flows;(2) network configuration in the system with return flows;(3) production planning with reuse of parts and materials;(4) vehicle routing problem with delivery and pickupsAlthough there are many quantitative models for reverse logistics,most of these studies focus on the single product system without considering product perishability. Manna & Chaudhuri(2006) and Chern et al(2008) pointed that perishablility of goods in stock is a very realistic feature.Therefore,it is necessary to incorporate product perishability into system hypothesizes.Hence,this dissertation focuses on the operational management issues about perishable product's reverse logistics.1.The scope of perishable product in this dissertation includes the obsolescent product and newsboy product.Due to the relative low value of deteriorating product' reverse logistics,the scope of perishable product in this dissertation just includes the obsolescent product and the newsboy product,excludes the deteriorating product,2.Based on the literature review,the topic of this dissertation is divided into two sub-topics involved into two activities in the whole reverse logistics process:the distribution problem involved in the used product collection activity and the inventory control problem involved in the reprocessing activity.Based on the review of relevant theories and literatures,this dissertation emphasizes on the perishable product inventory system with product return,the newsboy problem with resalable return and emergency order,the two products' inventory control system with product return and product substitution,the vehicle routing problem with simultaneously delivery and pickup,the vehicle routing problem with simultaneously delivery and pickup with the soft time windows,the vehicle routing problem with simultaneously delivery and pickup and hard time windows.The following contents are fulfilled:1.In this paper,using Discrete Event Simulation algorithm,inventory control model of the perishable product with product return is analyzed.The simulation results show the impact of product perishability on performance of the inventory system with product returns.According to the different manners how to deal with this unsatisfied demand, when the inventory on hand is not sufficient to meet all the customer's demand,three different inventory control model,that is,the inventory control model with product return with the assumption that the unsatisfied demand that can not be satisfied by its on-hand inventory,emergency order is placed to meet these unsatisfied,so shortage is not permitted;the inventory control model with product return and the assumption that the unsatisfied demand is lost and will not be satisfied in the future;the inventory control model with product return and the assumption that the unsatisfied demand will be backordered are presented for continuous review model under the "order up to" inventory control strategy.The algorithm proposed by Nelder and Mead is used to find the optimal value of the order-up-to parameters.According to many simulation curves and data,how the system serviceable inventory per time and the total system income change when some parameters are changed are analyzed.Finally,the perishable product system with product return system is compared with the traditional product system with product return and without considering product perishability.The results show the effect of product perishability on the optimal value of order-up-to parameter and system performance.The simulation results show that:if the uncertainty degree of product return is high,the quantity of outdated product in the perishable product inventory control system with reusable return is higher.With reusable product return,the total income of the perishale product inventory control system is higher than that of the traditional product system without considering product perishability.At the same time,the optimal 'order up to' parameter value of the perishable product inventory control system is lower than that of traditional inventory control system without considering product perishability2.In this paper,the newsboy problem with resalable returns is extended by incorporating the emergency order option.The total expected profit function is formulated and analyzed,and the optimal order quantity is derived.The results of computational experiment show that an emergency order can lead to a substantial increase in profit.The newsboy problem with resalable returns and emergency order is analyzed in this dissertation.From the view of the retailers,the profit function for the system with return and emergency order is formulated and analyzed.Meanwhile,the profit function for the system with returns and only once order opportunity is formulated and analyzed.According to the analysis of the two profit function,we give the formula to obtain the optimal order quantity.The comparison results shows that the expected profit of the newsboy system with resalable returns and emergency order is higher than the expected profit of the newsboy system with resalable returns and only once order opportunity,and its total order quantity of two order quantities is lower than the order quantity of the system with only once order opportunity.3.Based on the perishable high-tecch product one-way substitituion phenomenon and the reality that product substitution is a universal phenomenon,inventory optimization model of the two product system with product substitution and product rerun is developed in this dissertation.By the Markov Decision Process theory,the model is solved and gives stochastic optimal control strategy and algorithm.In this paper,the two product inventory models with constant demand,constant return or return with known discrete distribution are presented.Optimal control policy and control algorithm is derived by analyzing the quality of the system.4.In this paper,a new hybrid meta heuristics is provided for the vehicle routing problem with simultaneously delivery and pickups.The basic vehicle routing problem with simultaneously delivery and pickups is modified by including time windows constraints.The original meta-heuristics for the VRPSPD previously is revised to solve the extended VRPSPD models.The numerical experiments results show the validity of the meta heuristic presented in this dissertation.The new heuristic for the VRPSPD is a revised tour-partitioning algorithm. Using the order based on the net delivery demand and the minimum least vehicle number that is determined by the maximum value between the total customers' demand quantities and the total customers' pickup quantities and the vehicle maximum capacity,many basic tours are constructed as the initial input of the tour partitioning algorithm,instead of solving the corresponding travelling salesman problem to construct the basic tour.In the optimization procedure,the heuristic combines reactive tabu search with variable neighborhood.By the computational comparison with Nicola & Giovanni(2007) 's results,our heuristic is proved to be efficient to solve the VRPSPD problem.The results of numerical experiments show that:when the vehicle load capacity constraint is lower,the results found by our proposed algorithm are better than Nicola & Giovanni's(2007) results,from the view of either the total travel distance or the minmum vehicle number of the optimal solution.When the vehicle load capacity constraint is higher,the total travel distance of the optimal solution found by our proposed algorithm is longer than that of the Nicola & Giovanni(2007) proposed algorithm.But,the minmum vehicle number of the optimal solution found by our proposed algorithm is smaller than that of the Nicola & Giovanni(2007) proposed algorithm.At the same time,the time windows constraints that reflect the different instancy degree of businesses colletion activity due to the different life phases of the perishable products returned by different customer and the customer demand are incorporated into the basic VRPSPD problem.Some revisions of the new hybrid heuristic are done to be adaptive to solve the VRPSPD with soft time windows and the VRPSPD with hard time windows.