Learn More
The sample-path method is one of the most important tools in simulation-based optimization. The basic idea of the method is to approximate the expected simulation output by the average of sample observations with a common random number sequence. In this paper, we describe a new variant of Powell's UOBYQA (Unconstrained Optimization BY Quadratic(More)
DIRECT (DIviding RECTangles) is a deterministic global optimization algorithm for bound-constrained problems. The algorithm, based on a space-partitioning scheme, performs both global exploration and local exploitation. In this paper, we modify the deterministic DIRECT algorithm to handle noisy function optimization. We adopt a simple approach that(More)
We describe the application of a Bayesian variable-number sample-path (VNSP) optimization algorithm to yield a robust design for a floating sleeve antenna for hepatic microwave ablation. Finite element models are used to generate the electromagnetic (EM) field and thermal distribution in liver given a particular design. Dielectric properties of the tissue(More)
In many real-world optimization problems, the objective function may come from a simulation evaluation so that it is (a) subject to various levels of noise, (b) not differentiable, and (c) computationally hard to evaluate. In this paper, we modify Powell's UOBYQA algorithm to handle those real-world simulation problems. Our modifications apply Bayesian(More)
We investigate an on-line planning strategy for the fractionated radio-therapy planning problem, which incorporates the effects of day-today patient motion. On-line planning demonstrates significant improvement over off-line strategies in terms of reducing registration error, but it requires extra work in the replanning procedures, such as in the CT scans(More)
Principal Protected Absolute Return Barrier Notes (ARBNs) are structured products linked to an underlying security or an index. While these notes guarantee principal protection – return of face value – their upside potential is dependent on the level of the underlying security never falling outside of a predefined range. This, combined with the credit risk(More)
We investigate the use of optimization and data mining techniques for calibrating the input parameters to a discrete event simulation code. In the context of a breast-cancer epidemiology model we show how a hierarchical classifier can accurately predict those parameters that ensure the simulation replicates benchmark data within 95% confidence intervals. We(More)
In simulation-based optimization, we seek the optimal parameter settings that minimize or maximize certain performance measures of the simulation system. In this paper, we use a two-phase approach to calibrate simulation parameters using classification tools. This classification-based method is used in Phase I to facilitate the global search process and it(More)