Metaheuristic Approach to Assembly Line Balancing SUPAPORN SUWANNARONGSRI 1* and DEACHA PUANGDOWNREONG 2 1 Department of Industrial Engineering, Faculty of Engineering South-East Asia University 19/1 Petchkasem Road, Nongkham District, Bangkok, 10160, THAILAND 2 Department of Electrical Engineering, Faculty of Engineering South-East Asia University 19/1 Petchkasem Road, Nongkham District, Bangkok, 10160, THAILAND * corresponding author: supaporn-eng@sau.ac.th http://www.sau.ac.th Abstract: - In this paper, the metaheuristic method consisting of the adaptive tabu search (ATS) and the practicing heuristic (PH) technique is proposed to provide optimal solutions of the assembly line balancing (ALB) problems. With the multi-objective approach, the ATS is used to address the number of tasks assigned for each workstation, while the PH is conducted to assign the sequence of tasks for each workstation according to precedence constraints. The proposed approach is tested against six benchmark ALB problems suggested by Scholl and one real-world ALB problem. Comparisons of optimal results obtained by the proposed method with those obtained by the single objective approach are elaborated. Key-Words: - assembly line balancing, metaheuristic approach, adaptive tabu search, multi-objective function 1 Introduction Over six decades, the assembly line balancing (ALB) problem has been one of the most interesting topics among industrial researchers. Considered as the class of NP-hard combinatorial optimization problems [1], the ALB problem is one of the classic problems in industrial engineering. By literatures, several methods to provide the optimal solutions of the ALB problems were launched, for example, heuristic approaches [2-4], artificial intelligent (AI) search techniques such as the genetic algorithm (GA) [5] and the tabu search (TS) [6], and hybrid AI methods [7-10]. Based on the optimization context, the multi- objective optimizations can probably give better solutions than the single objective approach. With this idea, many researches have moved to use multi- objective approach to solve the ALB problems [11,12]. In 1989, Glover introduced the tabu search (TS) method to solve the combinatorial optimization problems [13,14]. Based on the neighborhood search approach, the TS method consists of two main strategies namely intensification and diversification [15,16]. In 2004, the modified version of the TS method named the adaptive tabu search (ATS) method was launched [17]. The ATS possesses two distinctive mechanisms denoted as back-tracking (BT) regarded as one of the diversification strategies and adaptive radius (AR) considered as one of the intensification strategies. The ATS can be regarded as one of the most powerful AI search techniques. Convergence proof and performance evaluation of the ATS have been reported [17,18]. The ATS has been widely applied to various real-world engineering problems, e.g. power system protection [19], dynamic system identification [20,21], control system design [22,23] and audio signal processing [24]. In 2008, the ATS associated with the partial random permutation (PRP) technique was developed to solve the ALB problems [25]. As previous results, it was found that such the approach could provide satisfy solutions. However, it spent amount of search time, when applied to solve the ALB problems containing a lot of tasks. In this paper, the metaheuristic approach consisting of the ATS and the practicing heuristic (PH) technique are proposed to provide optimal solutions of the ALB problems. The ATS is used to address the number of tasks assigned for each workstation, while the proposed practicing heuristic (PH) technique is conducted to arrange the sequence of tasks according to the precedent constraints. The workload variance, the idle time and the line efficiency are performed together as the multi- objective functions. To perform its effectiveness, the proposed approach is tested against six benchmark WSEAS TRANSACTIONS on SYSTEMS Supaporn Suwannarongsri, Deacha Puangdownreong ISSN: 1109-2777 200 Issue 2, Volume 8, February 2009