Two-stage meta-heuristic for part-packing and build-scheduling problem in parallel additive manufacturing

This paper studies the part-packing and build-scheduling problem in parallel additive manufacturing. The multi-part grouping, 2D placement, and 3D rotation should be determined in the part-packing problem. The build allocation and sequencing should be determined in the build-scheduling problem. The objective function is to minimize the makespan. A mixed integer linear programming (MILP) model is developed to find the optimal solution. Then, we propose a two-stage meta-heuristic that can decompose the proposed problem into the part-packing and the build-scheduling stages. In this paper, the two-stage meta-heuristic is applied to using a genetic algorithm (GA) and particle swarm optimization (PSO). To verify the performance of the two-stage meta-heuristic, two types of computational experiments are conducted. In the small-sized instance experiments, the two-stage meta-heuristic compares with the MILP model, and it shows a good performance. In the large-sized instance experiments, we compare two single-stage meta-heuristics applied to using the genetic algorithm and particle swarm optimization. The two-stage meta-heuristic shows the best performance among the proposed meta-heuristics. To analyze the experimental results and extract insights from the two-stage meta-heuristic, we conduct the robustness, impact of part rotation, and sensitivity analyses.

Automated summary

This paper investigates the part-packing and build-scheduling problem in parallel additive manufacturing. A mixed integer linear programming (MILP) model is developed to minimize the makespan. A two-stage meta-heuristic based on genetic algorithm (GA) and particle swarm optimization (PSO) is proposed and compared with other meta-heuristics. Experimental results demonstrate the effectiveness of the two-stage meta-heuristic.