Scheduling of build and post processes for decomposed parts in additive manufacturing

Abstract

Recently, production planning has been highlighted in the Additive Manufacturing (AM) literature. However, most AM studies have focused on a build process performed by an AM machine (also known as a 3D printer). Compared to the build process, AM studies for post-processing have lagged behind in production planning at the macro level. Post-processing needs highly sophisticated and complicated tasks requiring a lot of process time and cost. Therefore, some AM systems need to primarily cover post-processing as well as the build process, which leads to a more comprehensive approach to production planning. This paper addresses the AM systems based on flow-shop, comprehensively including the build and post processes. For post-processing, we consider trimming, coloring, and assembly processes. This paper introduces three scheduling policies: (1) all-for-one-job; (2) one-for-one-job; and (3) hybrid. The first policy is packing every part into the same job, which means that all parts are simultaneously fabricated during the same build process. The second policy is assigning each part to a particular job, which results in parts being fabricated one by one during the build process. The last policy is a compromise between the former two policies. We propose a mixed-integer programming (MIP) model to minimize the makespan for each policy. In computational experiments with test models, the scheduling optimization for AM is applied to various cases: multiple parts, scheduling policies, and multiple AM machines. One of the main comparisons provides that the MIP model of the hybrid policy is more preferred than the all-for-one-job policy, reducing the makespan by 27.68%. In conclusion, our paper quantitatively shows that packing is not a significant factor in minimizing the makespan when comprehensively considering build and post processes. It is a different perspective from that the traditional scheduling for AM has focused on a packing issue to improve the productivity in the build process. © 2022 Elsevier B.V.