Optimal sampling laws for stochastically constrained simulation optimization on finite sets

Optimal sampling laws for stochastically constrained simulation optimization on finite sets. Hunter, S. R. & Pasupathy, R. INFORMS Journal on Computing, 25(3):527–542, Summer, 2013. First Place, 2011 INFORMS Computing Society Student Paper Award.

Paper doi abstract bibtex 112 downloads

Consider the context of selecting an optimal system from among a finite set of competing systems, based on a stochastic objective function and subject to multiple stochastic constraints. In this context, we characterize the asymptotically optimal sample allocation that maximizes the rate at which the probability of false selection tends to zero. Since the optimal allocation is the result of a concave maximization problem, its solution is particularly easy to obtain in contexts where the underlying distributions are known or can be assumed. We provide a consistent estimator for the optimal allocation and a corresponding sequential algorithm fit for implementation. Various numerical examples demonstrate how the proposed allocation differs from competing algorithms.

@article{2013hunpas,
	Year = {2013},
	Author = {S. R. Hunter and R. Pasupathy},
	Title = {Optimal sampling laws for stochastically constrained simulation optimization on finite sets},
	Journal = {INFORMS Journal on Computing},
	shortjournal = {INFORMS J. Comput.},
	volume = {25},
	number = {3},
	month = {Summer},
	pages = {527--542},
	doi = {10.1287/ijoc.1120.0519},
	url_Paper = {http://web.ics.purdue.edu/~hunter63/PAPERS/2013hunpas.pdf},
	abstract = {Consider the context of selecting an optimal system from among a finite set of competing systems, based on a stochastic objective function and subject to multiple stochastic constraints. In this context, we characterize the asymptotically optimal sample allocation that maximizes the rate at which the probability of false selection tends to zero. Since the optimal allocation is the result of a concave maximization problem, its solution is particularly easy to obtain in contexts where the underlying distributions are known or can be assumed. We provide a consistent estimator for the optimal allocation and a corresponding sequential algorithm fit for implementation. Various numerical examples demonstrate how the proposed allocation differs from competing algorithms.},
	keywords = {simulation optimization > single-objective > stochastically constrained > ranking and selection},
	bibbase_note = {<span style="color: green">First Place, 2011 INFORMS Computing Society Student Paper Award.</span>}}

Downloads: 112