Integrated Production And Distribution Scheduling Under Make-to-Order And Direct Shipping Mode

Integrated production and distribution scheduling problem was studied for thecross-regional supply chain of custom-made or time-sensitive products. Decisions ofdifferent functions including order assignment, production scheduling and batchdelivery scheduling were integrated for a simultaneous optimization to increase theprofit, decrease the cost, and improve customer service level. In the problem,make-to-order mode and direct shipping mode were adopted in production stage anddistribution stage respectively, and finished orders are batch delivered to thecustomers immediately or shortly after production.Firstly, a mathematically model was built to minimize a weighted sum ofmaximum lead time and total cost of all orders, the optimal conditions and complexityof the problem were analyzed. To solve the problem with single plant and multiplecustomers, an exact algorithm was designed. In addition, a hybrid tabu searchalgorithm based on vector array representation was designed to solve the problemwith multiple plants and customers. In the above models, there is only one machine ineach plant. Models with parallel machines in each plant were also considered, twotabu search algorithms based on vector array representation were given to solve theproblem with single plant and multiple customers and the problem with multipleplants and customers, respectively. Experiment results and analysis demonstrate thevalidity of the designed algorithms and the vector array representation.Secondly, another mathematically model was built to minimize a weighted sumof total lead time and total cost of all orders, the optimal conditions and complexity ofthe problem were also analyzed. To solve the problem with single plant and singlecustomer, a dynamic programming method was designed. In addition, a hybrid tabusearch algorithm adopting adaptive selection was designed to solve the problem withmultiple plants and single customer. In the above models, there is only one machine ineach plant. Models with parallel machines in each plant were also studied, two hybridtabu search algorithms were designed to solve the problem with single plant andsingle customer and the problem with multiple plants and single customer,respectively. Experiment results verify the substantial advantage of the integratedapproach and confirm the validity of the designed algorithm and its adaptive selection. Thirdly, two mixed integer programming models were applied for the problemwith multiple parallel-machine plants and single customer to minimize the sum oftotal cost and to maximum the sum of total neat profit of all plants under the conditionthat all orders are finished before due date, respectively. Based on optimal conditionanalysis, two tabu search algorithms were designed to solve the models respectively.Experiment results verify the substantial advantage of the integrated approach andconfirm the validity of the designed algorithm.Finally, an absolute robust optimization model for the problem with multipleparallel-machine plants and single customer was designed to minimize a weightedsum of maximum lead time and total cost of all orders, in which the processing timeand processing cost of each order is uncertain. Experiment results verify thesubstantial advantage of the absolute robust optimization approach. In addition, thecircumstance that decision maker would underestimate task was also considered, theconclusion that underestimating tasks will make an evil influence on the system wasverified by experiment results, and an anchoring and adjustment heuristic method wasdesigned to reduce the evil influence. Experiment results confirm the validity of thedesigned method.