Proceedings of the 7 th Asia Pacific Industrial Engineering and Management Systems Conference 2006 17-20 December 2006, Bangkok, Thailand ________________________________________ † : Corresponding Author Optimization of Scheduling Method for Anthropocentric Manufacturing System Bandit Suksawat †, Weiming He, Hiroyuki Hiraoka and Tohru Ihara Department of Precision Mechanics, Faculty of Science and Engineering, Chuo University 1-13-27 Kasuka, Bumkyo-ku, Tokyo, 112-8551, JAPAN +813-3817-1822, Email: bandit@mail-ihara.mech.chuo-u.ac.jp,{hewm,hiraoka,ihara18}@mech.chuo-u.ac.jp Abstract. This study proposes an anthropocentric manufacturing system (AMS) scheduling method and also examines the optimum pattern for skill-based scheduling based on leader decision, using three scheduling patterns. These are: (i) fixed machine type, (ii) invariable machine type, machine model and machining method for some jobs, and (iii) constant machining method. The scheduling process was divided into three stages, including the application of a thought model to help leaders determine the degree of match between the job and the technicians’ characteristics, and the evaluation of the technician’s skill levels for each job in relation to the job characteristics. Using scheduling performance indexes, evaluation by a total earliness index (TEI), a total skill level index (TSI), and a man-hour index (MHI) the analysis shows that the scheduling pattern (iii) was the best pattern for this proposed scheduling method because at a constant value of TEI and MHI, the TSI was highest when compared with the two other patterns. With this method, the leader needs to only determine the machining method as selections of suitable machine types and machine models for each job and technician will be decided by the GA mechanism. Keywords: Anthropocentric Manufacturing System, Skill-based Scheduling, Genetic Algorithm, Optimization, Thought Model 1. INTRODUCTION Scheduling is the task of allocating available production resources (men, materials and machines) to jobs over time. A good scheduling strategy may help manufacturing enterprises respond to market demands quickly and to run plants efficiently. The modern manufacturing system, as a flexible manufacturing cell (FMC), is a group of machines working together to perform a set of functions on a particular part or product. A manufacturing cell can produce more than one family part as long as each part can be completely processed in that cell (Asokan et al. 2001). However, advanced technology does not improve productivity unless accompanied by a coherent labor strategy and operated by skilled technicians (Daude et al. 1998; Jensen 2001). The integration of people and technology in the factory according to an anthropocentric manufacturing system (AMS) is accepted as essential to improving productivity (Filip et al. 2002). AMS is a human-centered strategy that involves the combination of production technology, socio-technical production systems, and/or strategic production objectives. This strategy is based on the idea that human centered manufacturing could increase the productivity and operational efficiency of FMCs, which depend mainly on the experiences, and skills of technicians (Ito 1993; Nakazawa 1993, Rod 1994). However, it is difficult and complicated to allocate skilled technicians to suitable jobs because of the variety of technicians, job characteristics, machine types and machine models. This problem limits the application of manufacturing cell scheduling for the operational efficiency of AMS, even as manufacturing technologies in developing countries increasingly require skilled technicians to ensure that the technologies will function successfully and efficiently. AMS, therefore, could be an alternative to solve that problem and improve the productivity and operational efficiency in FMC. In recent decades, many researchers have used genetic algorithms (GA) to solve flexible manufacturing scheduling problems with an optimal constrain. GA has been used to schedule jobs in a sequence dependent setup environment for a minimal loss of time. With this method, all jobs are scheduled on a single machine; each job has a processing time and a due date (Jawahar et al. 1998). It is also used to schedule jobs in non-sequence dependent setup environments. The jobs are scheduled on one machine with 1634