A HYBRID GENETIC ALGORITHM FOR BALANCING ASSEMBLY LINES WITH COMPATIBILITY CONSTRAINTS
OCTAV BRUDARU 1, DIANA POPOVICI 2, CINTIA COPACEANU 3 1. Institute of Computer Science, Romanian Academy, Jassy Subsidiary “Gh. Asachi” Technical University Jassy, Department of Management and Production Systems Engineering, e-mail: brudaru@tuiasi.ro
2. Institute of Computer Science, Romanian Academy, Jassy Subsidiary
3. “Gh. Asachi” Technical University Jassy, Department of Management and Production Systems Engineering
The paper presents a hybrid genetic algorithm for deterministic assembly line balancing (ALB) problem with a single model and an additional constraint, which requires that the workstations are compatible with a given cover of the assembly tasks. The performance criteria are the minimizing of the idle time and the smoothing index. The algorithm includes a special procedure to generate the cover sets and a special mutation operator preserving the topological order. It is also combined with an efficient greedy procedure proper to the problem. All genetic operators are applied with dynamic probabilities that favour the creating and preserving of good constructive blocks. The experimental investigation proves the ability of the hybrid method to find good solutions to this type of balancing problem.
