SIMULATED ANNEALING FOR RESOURCE-CONSTRAINED SCHEDULING

We describe an application of several procedures to a resource-constrained scheduling problem. Basic and augmented neighborhood search procedures are first developed, using a unique way to generate neighborhoods. We then present a simulated annealing procedure which can be viewed as an evolution of the neighborhood search procedures. Computational results indicate that the procedures grow in power as they evolve, with the simulated annealing procedure providing the best results, and further, that it is a viable approach for a very difficult scheduling problem.