Simulation-based scheduling package-models and solutions

This paper describes methods of modeling and solutions of the simulation-based scheduling package which gives "good" feasible schedules for production systems by systematic search using the simulator. In the simulator, production systems are modeled by generalized Petri nets, and both forward and backward simulations can be carried out. These two alternatives offer us an earliest start schedule and a latest start schedule, respectively. As methods of making systematic search, a branch-and-bound method and a heuristic method using priority rules have been incorporated. In this paper, three kinds of new ways of finding an approximately optimal schedule based on the branch-and-bound method and two kinds of search methods, i.e. simulated annealing methods and genetic algorithms, are presented. It is assured by numerical experiments that the package can give reasonable solutions.<
>