Response Surface Methodology as a Sensitivity Tool in Decision Analysis.
Abstract
The purpose of this study is to evaluate response surface methodology as a sensitivity analysis tool in the area of decision analysis. The advent of low-cost personal computer software, such as DPLTM, has created an accessible tool with the ability to frame and solve influence diagrams for decision problems. This study provides a comparison of current sensitivity analysis techniques vs those made possible through response surface methodology (RSM). Sensitivity analysis alternatives are demonstrated on a decision problem concerning the evaluation of force structure options for the Department of Defense. Sensitivity analysis is performed on both one-way and two-way perturbations of input variables. The most significant contribution of this thesis is the effectiveness of RSM techniques for the sensitivity analysis of a decision problem. RSM is proven to be an effective and highly efficient approach. The combination of two analysis tools, the influence diagram solver and the RSM sensitivity analysis, has effectively saved the decision-maker valuable resources and increased the information made available. This capability will enhance the analysis of uncertainty in decision problems.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1995
- Accession Number
- ADA293910
Entities
People
- David A. Meyers
Organizations
- Air Force Institute of Technology