Genetic Algorithms for an Optimal Line Balancing Problem with Workers of Different Skill Levels

Abstract

This paper discusses a new combinatorial optimization problem which occurs in line balancing for real assembly lines demanding skilled operations. On the contrast with conventional assembly lines such as automotive in which each operation is associated with a standard processing time, it is assumed that each operation time depends on assigned worker's skill and there exists an upper bound on the number of operations to be assigned to each worker. Three genetic algorithms (GAs) which have different genotypes and different decoding procedures are discussed for this problem. The genotype in the first GA is expressed by sequencing the operation numbers, and an effective heuristic rule is introduced into the decoding procedure. In the second GA, the genotype is expressed by sequencing the sets of operations to be assigned to each worker. In the third GA, the genotype is expressed by sequencing the worker numbers executing each operation in the order of operation numbers. These GAs are compared by numerical experiment based on real conditions.