Confidence Interval Estimation for Simulation Response Surfaces
Abstract
In this paper, the author continues the investigation of response surface methodology in simulation, begun in previous papers. Suppose that g is a linear output function of an input parameter lambda, over some range a < or = lambda < or = b, and that observations are taken at two parameter settings lambda sub 1 and lambda sub 2. It is shown that if certain conditions are satisfied, then a confidence interval may be obtained for g(lambda) at any point a < or = lambda b without making any additional observations. An illistration is provided which makes use of the regenerative approach for analyzing simulations.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1974
- Accession Number
- ADA003901
Entities
People
- Michael A. Crane