Reducing and Calibrating for Input Model Bias in Computer Simulation

Morgan, Lucy and Rhodes-Leader, Luke and Barton, Russell (2022) Reducing and Calibrating for Input Model Bias in Computer Simulation. INFORMS Journal on Computing. p. 1. ISSN 1091-9856

[img]
Text (Reducing and Calibrating for Input Model Bias - pre-print)
Reducing_and_Calibrating_for_Input_Model_Bias_pre_print.pdf - Submitted Version
Available under License Creative Commons Attribution-NonCommercial.

Download (939kB)

Abstract

Input model bias is the bias found in the output performance measures of a simulation model caused by estimating the input distributions/ processes used to drive it. To be specific, when input models are estimated from a finite amount of real-world data they contain error and this error propagates through the simulation to the outputs under study. When the simulation response is a non-linear function of its inputs, as is usually the case when simulating complex systems, input modelling bias is one of the errors to arise. In this paper we introduce a method that re-calibrates the input parameters of parametric input models to reduce the bias in the simulation output. The method is shown to be successful in reducing input modelling bias and the total mean squared error caused by input modelling.

Item Type:
Journal Article
Journal or Publication Title:
INFORMS Journal on Computing
Uncontrolled Keywords:
/dk/atira/pure/subjectarea/asjc/1700/1706
Subjects:
ID Code:
168258
Deposited By:
Deposited On:
11 Apr 2022 11:00
Refereed?:
Yes
Published?:
Published
Last Modified:
21 May 2022 00:36