Abstract
The purpose of this study is to solve the Preventive Maintenance(PM) scheduling problem on a CNC machine. The machine breaks down will have additional cost to repair it and insert a PM can decrease the cost.Our purpose is to find the best solution that is defined by the shortest total completion time. Each solution is characterized by different machine rate,scheduling and PM decisions.Higher machine rate will have higher probability to break down so we need insert a PM to decrease the probability. In our research, the work process time, probability of break down, Repair time and setup time are stochastic variables so we use Optimization via Simulation to solve this problem. So I combine the Modified Heuristic algorithm with Ranking and Selection(R＆S) to find the best solution. Finally, we found combining the R＆S with GA or GA + SA can improve and Stabilize the quality of the best solution. When Facing the situation of complicated problems, heuristic algorithms have to run more generation to decide the best solution.It may cost many samples and time, but combing R＆S can decrease the samples and time. We can apply VRT on the procedure to decrease the samples and time or use this procedure on more complicated problems in the future.