Design of a footwear assembly line using dimulation-based ALNS

Competitive advantages of footwear manufacturing companies today lie in the ability to improve the efficiency and effectiveness of resource utilization through constantly eliminating wastes. This paper presents an innovative approach to designing a footwear assembly line with uncertain task times and parallel workstations. The measure defined to characterize the performance of the assembly line is Pair/Person/Hour (PPH). Two operational parameters, namely buffer size and the number of resources are taken into account. The objective is to maximize the performance measure of the assembly line by determining an optimized setting of the operational parameters and optimal task assignment. The solution approach utilizes Discrete Event Simulation (DES) to model the performance measure under consideration of process constraints and variability. An Adaptive Large Neighborhood Search (ALNS) heuristic is next developed and integrated into the simulation model to search for problem domain. A case study of an industrial footwear manufacturing factory is conducted to demonstrate the effectiveness of the proposed approach.