Designing Machine Layout Using Tabu Search and Simulated Annealing

Abstract

The machine layout design (MLD) problem usually arises when a manufacturing company aimed to expand their production capacity and/or decrease the handling distances of materials or parts flow through a predefined sequence of machines for manufacturing a product. The problem is known to be Non-deterministic Polynomial (NP) hard, which is usually solved by metaheuristics such as Simulated Annealing (SA) and Tabu Search (TS). This paper presents the application of SA and TS for minimising the material handling distance associated with the layout required for manufacturing process of multiple products. A computer based machine layout designed tool was developed and tested using five datasets adopted from literature. The analysis on the computational results obtained from numerical experiments indicated that the average best so far solutions obtained from SA are marginally better than TS but the average execution times taken by TS were at least 50% faster than SA. The convergences of best so far solutions during TS iteration were quicker than those obtained from SA especially for small- and medium-size problems.