Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Simulation and performance sensitivity analysis of lot scheduling policies

Thomas Theodoros

Full record

---

Summary

We consider a production system making different type of products in batches to avoid frequent product changes and set ups, which incur cost and delays. The system produces to satisfy orders having random arrival times, random volumes and varieties of products. A production policy is a decision rule which sets the batch production start and completion times of every product when all of them are produced by the same machine. Using simulation, a system's production operational scenario can be generated under any given production policy and collect statistics of important economic indicators such as production cost, inventory holding cost and delays in filling customer orders. Finding an optimal policy is based on the sensitivity analysis which calculates how specific small changes in a policy (e.g imposing a short delay in the start or completion time of some product type) would affect the total cost. Such information usually requires additional simulations to be carried out using all possible modified policies. This work develops a Markov model in order to describe the operation of the system and simultaneously applies the, so called, standard clock technique, which enables a sensitivity analysis to be performed during a single simulation involving only the original policy at a small additional computational cost. A number of numerical experiments are carried out to investigate the accuracy of the proposed method and its speed against a standard sensitivity analysis that requires several simulation runs.

Available Files

Services