OBJECT-ORIENTED JOB-SHOP SCHEDULING USING GENETIC ALGORITHMS

An object-oriented scheduling system for a job-shop manufacturing cell has been developed and tested. A powerful optimization technique base d on genetic algorithms was used to improve the routing of subassembly parts in the manufacturing cell. The scheduling system is composed of three interactive modules: the due date algorithms, the genetic algor ithms, and the manufacturing cell simulator. These modules exchange in formation to accomplish 'genetic optimization' and to fulfill time-dep endent and technological constraints simultaneously, using a composite objective function. The due date algorithms use the production sequen ces and processing times of individual parts of a single product to ge nerate production priorities. The genetic algorithms module creates a population of feasible machine routings and schedules, and passes it o n to the manufacturing cell simulator. The manufacturing cell simulato r performs the production plans under different scheduling configurati ons. The genetic algorithms module uses the fitness value of each sche duling alternative progressively to improve the production schedule un til a convergence criterion is satisfied. (C) 1997 Elsevier Science B. V.