Large deviations of uniformly recurrent Markov additive processes
Advances in Applied Mathematics
The asymptotic efficiency of simulation estimators
Operations Research
Rates of convergence of ordinal comparison for dependent discrete event dynamic systems
Journal of Optimization Theory and Applications
Simulation Budget Allocation for Further Enhancing theEfficiency of Ordinal Optimization
Discrete Event Dynamic Systems
A large deviations perspective on ordinal optimization
WSC '04 Proceedings of the 36th conference on Winter simulation
A new perspective on feasibility determination
Proceedings of the 40th Conference on Winter Simulation
A large deviations view of asymptotic efficiency for simulation estimators
Proceedings of the 40th Conference on Winter Simulation
Large-deviation sampling laws for constrained simulation optimization on finite sets
Proceedings of the Winter Simulation Conference
Hi-index | 0.00 |
Consider a simulation estimator α(c) based on expending c units of computer time to estimate a quantity α. In comparing competing estimators for α, a natural figure of merit is to choose the estimator that minimizes the computation time needed to reduce the error probability P(|α(c) − α| ε) to below some prescribed value δ. In this paper, we develop large deviations results that provide approximations to the computational budget necessary to reduce the error probability to below δ when δ is small. This approximation depends critically on both the distribution of the estimator itself and that of the random amount of computer time required to generate the estimator, and leads to different conclusions regarding the choice of preferred estimator than those obtained when one requires the error tolerance ε to be small. The “small ε” regime leads to variance-based selection criteria, and has a long history in the simulation literature going back to Hammersley and Handscomb.