The knowledge gradient algorithm using locally parametric approximations

Bolong Cheng, Arta A. Jamshidi, Warren Buckler Powell

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We are interested in maximizing a general (but continuous) function where observations are noisy and may be expensive. We derive a knowledge gradient policy, which chooses measurements which maximize the expected value of information, while using a locally parametric belief model which uses linear approximations around regions of the function, known as clouds. The method, called DC-RBF (Dirichlet Clouds with Radial Basis Functions) is well suited to recursive estimation, and uses a compact representation of the function which avoids storing the entire history. Our technique allows for correlated beliefs within adjacent subsets of the alternatives and does not pose any a priori assumption on the global shape of the underlying function. Experimental work suggests that the method adapts to a range of arbitrary, continuous functions, and appears to reliably find the optimal solution.

Original languageEnglish (US)
Title of host publicationProceedings of the 2013 Winter Simulation Conference - Simulation
Subtitle of host publicationMaking Decisions in a Complex World, WSC 2013
Pages856-867
Number of pages12
DOIs
StatePublished - 2013
Event2013 43rd Winter Simulation Conference - Simulation: Making Decisions in a Complex World, WSC 2013 - Washington, DC, United States
Duration: Dec 8 2013Dec 11 2013

Publication series

NameProceedings of the 2013 Winter Simulation Conference - Simulation: Making Decisions in a Complex World, WSC 2013

Other

Other2013 43rd Winter Simulation Conference - Simulation: Making Decisions in a Complex World, WSC 2013
CountryUnited States
CityWashington, DC
Period12/8/1312/11/13

All Science Journal Classification (ASJC) codes

  • Modeling and Simulation

Fingerprint Dive into the research topics of 'The knowledge gradient algorithm using locally parametric approximations'. Together they form a unique fingerprint.

  • Cite this

    Cheng, B., Jamshidi, A. A., & Powell, W. B. (2013). The knowledge gradient algorithm using locally parametric approximations. In Proceedings of the 2013 Winter Simulation Conference - Simulation: Making Decisions in a Complex World, WSC 2013 (pp. 856-867). [6721477] (Proceedings of the 2013 Winter Simulation Conference - Simulation: Making Decisions in a Complex World, WSC 2013). https://doi.org/10.1109/WSC.2013.6721477