A Genetic Algorithm for Assembly Sequence Planning

Abstract

This work presents a genetic algorithm for assembly sequence planning. This problem is more difficult than other sequencing problems that have already been tackled with success using these techniques, such as the classic Traveling Salesperson Problem (TSP) or the Job Shop Scheduling Problem (JSSP). It not only involves the arranging of tasks, as in those problems, but also the selection of them from a set of alternative operations. Two families of genetic operators have been used for searching the whole solution space. The first includes operators that search for new sequences locally in a predetermined assembly plan, that of parent chromosomes. The other family of operators introduces new tasks in the solution, replacing others to maintain the validity of chromosomes, and it is intended to search for sequences in other assembly plans. Furthermore, some problem-based heuristics have been used for generating the individuals in the population.