A Genetic Algorithm Applied to Resource Management in Production Systems

This chapter presents an Evolutionary Computation approach to scheduling in industrial environments where schedule cost is related to resource management. To construct production schedules with near-optimal resource exploitation, a genetic-algorithm-based system was developed, utilising a direct string representation of schedules, problem-specific initialisation and recombination procedures, elitism and local improvement of solutions. The scheduler was originally evaluated on a problem of optimal energy consumption in a textile factory. It was subsequently upgraded and installed as a resource management tool in the ship repair division of a shipyard, where large numbers of activities have to be scheduled so that optimal work load is provided for workers of different trades. Employing the scheduler, significant savings in resource management were made possible in both production systems.