Package diviner

Provides functionality for measuring the level of promise of a design candidates.

See:
Description

Interface Summary

BoundConstrainedDiviner	Computes a figure of merit for sampling at some point, based on information obtained from an ApproxModel.
InequalityConstrainedDiviner	Diviner for inequality-constrained problems.
MultiObjectiveDiviner	A diviner that uses multiple objectives as measures of promise.

Class Summary

BasicBoundConstrainedDiviner	Implements basic functionality for BoundConstrainedDiviner.
BasicInequalityConstrainedDiviner	Basic diviner for inequality-constrained problems.
BasicMultiObjectiveDiviner
ConstrainedEIDiviner	Constrained version of the EIDiviner.
ConstrainedImprovementDiviner	ImprovementDiviner for constrained problems.
ConstrainedMultiMetricDiviner	Multi-metric diviner for constrained problems.
EIDiviner	Implements an expected improvement BoundConstrainedDiviner.
ImprovementDiviner	Focuses on various ways of divining merit using the posterior distribution at a point and some computation of improvement.
MultiMetricDiviner	Uses multiple figures of merit to compute 'goodness' of a point for sampling.
PIDiviner	Implements a diviner based on probability of improvement.
RiskDiviner	Another way of trading performance and risk: quantify the risk as the probability of not achieving a given performance.
RiskPerfTradeDiviner	BoundConstrainedDiviner that returns mean - k*stdDev.
RiskPerfTypeIDiviner	Multiobjective diviner that uses mu and sigma as objectives to be simultaneously minimized.
RiskPerfTypeIIDiviner	Multiobjective diviner that uses mu - k*sigma and sigma as objectives to be simultaneously minimized.
RiskPerfTypeIIIDiviner	Multiobjective diviner that uses mu - k*sigma and mu as objectives to be simultaneously minimized.
RiskPerfTypeIVDiviner	Multiobjective diviner that uses mu and -sigma as objectives to be simultaneously minimized.
RiskPerfTypeVDiviner	Multiobjective diviner that uses mu - k*sigma and mu + k*sigma as objectives to be simultaneously minimized.

Package diviner Description

Provides functionality for measuring the level of promise of a design candidates. For lack of a better small word to describe an entity that computes the promise of yield at some unknown point, the word diviner was chosen. Examples of such metrics include expected improvement, probability of improvement, etc.