Simulation optimization: methods and applications
Proceedings of the 29th conference on Winter simulation
| Hi-index | 0.00 |
Discrete-event simulation is widely used to analyse and improve the performance of manufacturing systems. The related optimization problem often includes integer design variables and is defined by objective function and constraints that are expected values of stochastic functions. These stochastic functions have to be evaluated via the simulation model at the discrete levels of the integer design parameters. For such a simulation optimization problem with integer variables we have developed an optimization strategy that is based on a series of linear approximate subproblems. Each subproblem is built from the outcomes of simulation experiments. A D-optimal design of experiments is used to plan the simulation experiments. Stochasticity in constraint and objective functions is dealt with explicitly using safety indices. Two test problems will be presented to illustrate the optimization strategy. This includes a simulation based four-station production flow line problem.