Synchronized production scheduling for order fulfillment

Abstract

Currently, manufacturing companies are under pressure of satisfying punctuality of customer orders. Besides, in order fabrication, jobs from a customer are often separately processed in dispersed manufacturing resources, such as different machines, facilities, or factories. This leads to difficulties of processing customer orders in a simultaneous manner. This thesis proposes the concept of manufacturing synchronization (MfgSync) in terms of simultaneity and punctuality, and systematically studies the scheduling level of MfgSync for order fulfilment. Four typical scenarios related to MfgSync at the scheduling level are investigated, which are referred to as synchronized production scheduling (SPS) problems. The first scenario investigates SPS in an assembly flowshop system motivated by a real-life company. A measure named mean longest waiting duration (MLWD) is proposed for simultaneity. The mean earliness and tardiness is adopted for punctuality. A scheduling model is developed considering both simultaneity and punctuality using the linear weighted sum method. A modified genetic algorithm (GA) is then proposed. Numerical studies demonstrate the value of simultaneity on reducing finished products inventory. Simultaneity is greatly affected by the production system configuration, especially in peak seasons. Moreover, the weights of simultaneity and punctuality are suggested. The second scenario discusses SPS considering dynamic arrival orders in a hybrid flowshop (HFS). Periodic policy is employed to decompose the problem into a series of static sub-problems. A base model for each sub-problem is formulated by mixed integer programming (MIP) to minimize MLWD subject to punctuality constraint. A GA is proposed to address the base model. Numerical studies show that the proposed approach performs better than the online scheduling algorithms. A larger decision interval is preferred in off season than peak season. It is easier to obtain simultaneity when bottleneck is located at the entrance than at the middle, followed by at the exit. Serious bottleneck effect leads to bad simultaneity in HFS. The third scenario studies synchronized scheduling of production and shipping with individual decision autonomy. A novel bilevel scheduling model using MIP is proposed. A bilevel heuristic algorithm (BHA) based on simulating annealing are correspondingly proposed. Comparisons with the traditional sequential and integrated approaches verify the effectiveness and efficiency of the proposed approach. Furthermore, experimental results show that shipping capacity has a considerable impact on overall makespan reduction. This study also investigates the allocation of fixed shipping capacity onto shipping blocks and finds that large more (LM) strategy outperforms average allocation (AA) strategy and large less (LL) strategy. The fourth scenario studies SPS in cloud manufacturing (CMfg) with a new client coordination mode. The success of the coordination depends on whether every participant can be better off than to act independently. A cooperative game is built to analyze the client coordination. Then a core incentive payment mechanism is proposed to facilitate a grand coalition. Service booking and scheduling are tackled as an integrated optimization problem to generate coordinated solutions for clients. Based on dynamical programming, two optimal solution algorithms are correspondingly proposed. Computational results show that the proposed approach can make substantial cost savings for clients.