Title :
Response Gradient Estimation using Mathematical Programming Models of Discrete-Event System Sample Paths
Author :
Chan, Wai Kin Victor ; Schruben, Lee W.
Author_Institution :
Dept. of Decision Sci. & Eng. Syst., Rensselaer Polytech. Inst., Troy, NY
Abstract :
This paper illustrates the use of mathematical programming in computing gradient estimators. Consistency property of these estimators is established under the usual assumptions for IPA gradient estimator consistency. A finite difference tolerance limit is introduced. For complex discrete-event systems, more concise linear programming representations are developed. These new representations provide a direct way of calculating gradient estimates
Keywords :
discrete event systems; gradient methods; mathematical programming; discrete-event system sample paths; finite difference tolerance; linear programming; mathematical programming models; response gradient estimation; Algorithm design and analysis; Analytical models; Computational modeling; Discrete event simulation; Discrete event systems; Dynamic programming; Finite difference methods; Linear programming; Mathematical model; Mathematical programming;
Conference_Titel :
Simulation Conference, 2006. WSC 06. Proceedings of the Winter
Conference_Location :
Monterey, CA
Print_ISBN :
1-4244-0500-9
Electronic_ISBN :
1-4244-0501-7
DOI :
10.1109/WSC.2006.323083
