Meta-heuristics from nature for the loop layout design problem

This paper addresses the unidirectional LLDP using a differential evolution algorithm (DEA); a modern meta-heuristic from the field of evolutionary computation that is known to be NP-hard and compared to other known meta- heuristics such as genetic algorithms and simulated annealing.

Abstract

The loop-layout design problem (LLDP) arises when the machines in a flexible manufacturing system (FMS) are arranged in a closed ring-like network and the materials are transported around this network in only one direction. Evaluation of this layout is usually performed by estimating the traffic congestion, i.e., the number of cycles spent by each part in the network until its processing through the required machines is completed. The problem is known to be NP-hard and thus the right way to proceed is through the use of heuristics techniques. This paper addresses the unidirectional LLDP using a differential evolution algorithm (DEA); a modern meta-heuristic from the field of evolutionary computation. The performance of the DEA is measured through multiple characteristic experiments and compared to that of other known meta-heuristics such as genetic algorithms and simulated annealing.