A Hybrid GA-CP Approach for Production Scheduling

Constraint programming (CP) uses constraints present in production scheduling problems to derive feasible schedules but encounters computational difficulties when problem complexity increases. A hybrid method is proposed which uses genetic algorithms for global search and CP for constraint solving. Its performance is illustrated with an example from precast production scheduling presented as constrained precast scheduling model (CPSM). Results show that the hybrid method is able to extend the range of schedule periods up to 30 days where a feasible solution can be found and returns better solutions than CP.