An evolutionary approach for the reconfiguration of an assembly-line system

Abstract

Market competition forces an assembly line system to enhance the degree of manufacturing flexibility to cope with the ever changing production requirements. The design of an assembly line system becomes difficult. Those difficulties have been investigated in the research in terms of "unknown", "uncertainty" and "unforeseeable". Recently, many reconfigurable conveyor-components have been developed for reconfiguring an assembly line system to tackle the "uncertainty" and "unforeseeable" problems. The components have different transporting paths, shapes, sizes, etc. As a result, system design process often faces a larger number of design alternatives; many possible system configurations can be formed to meet the various production needs. It is very difficult to select an appropriate configuration. This research aims at investigating artificial intelligence (AI) based methods to configure reconfigurable conveyor-components for a flexible assembly line system (FALS) to cope with ever changing production requirements. In order to achieve the aim of the research, the reconfigurable conveyor-components have been investigated. In fact, the different shapes and transporting motions of those reconfigurable conveyor-components are constructed by two basic movements i.e. linear and angular. They are general transporting properties of linear and rotatory conveyors. Thus the two conveyors are commonly utilized to reconfigure a flexible conveyor system for a FALS. In this research, the two conveyors have been selected to develop genetic algorithm (GA) for resolving the difficulty in a FALS reconfiguration. The GA simulates the evolution of an individual population based on the rule of the survival of the fittest for the given environment. Each individual in the population is called chromosome that is major data in the program. The transporting location, motions, and sizes of the linear and rotatory conveyors are encoded into binary string as chromosome to represent an assembly line layout for evaluation and evolution. The evaluation consists of the defined design criteria that are used to analyze the performance of each layout for evolution. In the evolution, the linear and angular motions of the conveyors are evolved to construct bended motion to generate diversified shapes of conveyor-components for reconfiguring an appropriate layout of FALS. Finally, in order to evaluate the performance of the GA approach, it is compared with simulated annealing (SA), and knowledge-based system (KBS). The three approaches are applied on the same reconfiguration problems and compared to the performance of the qualified layout identified by each approach and the computation times of the approaches. The comparison shows that the proposed GA performs better in all the performance measures than other approaches. However, the convergence rate of the GA is longer than KBS because the GA searches for global optimal solutions with more iterations.

Date of Award	15 Jul 2003
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Kin Lim John HO (Supervisor)