Genetic algorithms in integrated process planning and scheduling NORHASHIMAH MORAD 1 * and AMS ZALZALA 2 1 School of Industrial Technology, Universiti Sains Malaysia, 11800 Minden, Penang, Malaysia 2 Department of Automatic Control and Systems Engineering, University of Sheeld, Mappin Street, Sheeld, S1 3JD, UK Process planning and scheduling are actually interrelated and should be solved simulta- neously. Most integrated process planning and scheduling methods only consider the time aspects of the alternative machines when constructing schedules. The initial part of this paper describes a genetic algorithm (GA) based algorithm that only considers the time aspect of the alternative machines. The scope of consideration is then further extended to include the processing capabilities of alternative machines, with dierent tolerance limits and processing costs. In the proposed method based on GAs, the processing capabilities of the machines, including processing costs as well as number of rejects produced in alternative machine are considered simultaneously with the scheduling of jobs. The formulation is based on multi- objective weighted-sums optimization, which are to minimize makespan, to minimize total rejects produced and to minimize the total cost of production. A comparison is done with the traditional sequential method and the multi-objective genetic algorithm (MOGA) approach, based on the Pareto optimal concept. Keywords: Integrated process planning and scheduling, multi-objective genetic algorithms 1. Introduction Manufacturing systems involve a lot of problems that are actually interrelate and should be solved concurrently. The traditional method of engineering design is done sequen- tially and known as serial engineering. Traditionally, the machine to process a certain operation is chosen based on the unit cost of operation. While this method allows easier scheduling, it could cause some machines to be overloaded and thus create bottlenecks. Consequently, more jobs can become tardy. This approach unnecessarily restricts the capability of manufacturing cells, where most operations can be processed in more than one machine in a cell. In practice, scheduling and planning problems rarely in- volve only a single consideration as manifested in classic combinatorial problems (Zentner et al., 1994). These prob- lems involve multiple objectives that need to be addressed simultaneously. Hence, the actual optimization problem is to determine the process plan and schedule concurrently. This paper describes the integration of process planning and scheduling using genetic algorithms (GAs), founded upon the principle of evolution (Goldberg, 1989; Mi- chalewicz, 1994). The scheduling of jobs in manufacturing cells is similar to the job-shop problem (JSP). However, this type of scheduling involves more ¯exibility because operation of the part can be performed in alternative ma- chines that have dierent processing times. Due to this ¯exibility, solution is greatly increased. The scheduling problem is than extended to include process planning as well. The scope of consideration of the alternative machines include dierent capabilities and cost to operate, as well as dierent processing times. The scheduling problem is then formulated as a multi-objective problem with the objectives of minimizing makespan, minimizing total number of rejects and minimizing total processing cost. These criteria will simultaneously optimize process planning and scheduling of the parts. The remainder of the paper is arranged as follows: Sec- tion 2 describes the scheduling problem in manufacturing cells, followed by process planning in Section 3. Section 4 describes the need for an integrated approach. Formula- tions based on GAs for the scheduling problem are given in Section 5, and for the integrated process planning and scheduling problem in Section 6. Finally, some discussion and the conclusions are given in Section 7. 2. Scheduling in manufacturing cells Similar to the ¯exible manufacturing systems (FMS), manufacturing cells may consist of Computer Numerical *Author to whom correspondence should be addressed. Journal of Intelligent Manufacturing (1999) 10, 169±179 0956-5515 Ó 1999 Kluwer Academic Publishers