Optimization of manufacturing system simulations using perturbation analysis and SENSE

Authors:
Rajan Suri;Gregory W. Diehl;Yu-Chi Ho
Affiliations:
University of Wisconsin, Dept. of Industrial Engrg., Madison, WI;Network Dynamics Inc., 1218 Mass Ave., Cambridge, MA;Harvard University, Divn. of Applied Sciences, Cambridge, MA
Venue:
WSC '85 Proceedings of the 17th conference on Winter simulation
Year:
1985

Citing 3
Cited 2

Perturbation analysis gives strongly consistent sensitivity estimates for the M/G/1 queue

Management Science
MANUPLAN: a precursor to simulation for complex manufacturing systems

WSC '85 Proceedings of the 17th conference on Winter simulation
The phantom customer and marked customer methods for optimization of closed queueing networks with blocking and general service times

SIGMETRICS '83 Proceedings of the 1983 ACM SIGMETRICS conference on Measurement and modeling of computer systems

MANUPLAN: a precursor to simulation for complex manufacturing systems

WSC '85 Proceedings of the 17th conference on Winter simulation
Testing strategies for simulation optimization

WSC '87 Proceedings of the 19th conference on Winter simulation

Quantified Score

Hi-index	0.00

Visualization

Abstract

Perturbation Analysis (P/A) of discrete event systems is a recent technique which enables sensitivity analysis of decision parameters from a single simulation run. Thus, from a single simulation of a manufacturing design, one can estimate the effects of changing many different decisions such as adding more machines or fixtures, changing buffer sizes or changing lot sizes. The first part of this paper gives an overview of this new technique. The second part describes SENSE (SENSitivity Evaluator), a software package implementing the theories of P/A in easily usable form. Application of SENSE to GPSS/H simulations is described, with examples of its use for the design of flow lines at a major U.S. corporation. By using SENSE a simulation team is able to improve their designs in less time, and with less manpower and CPU-time, than previously required. The result is better designs and better turnaround time for simulation activities in general.