Research Article
fDRIT- An Evaluation Tool for Transient Removal Methods in Discrete Event Stochastic Simulations
@INPROCEEDINGS{10.4108/eai.25-10-2016.2266603, author={Sushma Nagaraj and Armin Zimmermann}, title={fDRIT- An Evaluation Tool for Transient Removal Methods in Discrete Event Stochastic Simulations}, proceedings={10th EAI International Conference on Performance Evaluation Methodologies and Tools}, publisher={ACM}, proceedings_a={VALUETOOLS}, year={2017}, month={5}, keywords={initial transient detection start-up bias online methods for transient removal}, doi={10.4108/eai.25-10-2016.2266603} }- Sushma Nagaraj
Armin Zimmermann
Year: 2017
fDRIT- An Evaluation Tool for Transient Removal Methods in Discrete Event Stochastic Simulations
VALUETOOLS
ACM
DOI: 10.4108/eai.25-10-2016.2266603
Abstract
The choice of an initial state in a stochastic simulation often leads to a transient bias in the estimation results. A well-known solution for this problem is to detect and estimate this initial phase and its subsequent removal. A significant number of algorithms has been proposed in the literature for this task. Because of their heuristic nature and the stochastic experiment underlying each individual simulation run, their quality cannot be easily evaluated for a new application model or simulation tool.
The goal of this paper is a software framework for the systematic comparison of such algorithms to let tool developers evaluate and compare them for their specific needs in a fair and generic manner, and to allow researchers in the field of initial transient removal to compare and test their ideas. Along the way, a set of possible quantitative evaluation criteria is proposed. The framework is then used to evaluate and compare numerous existing algorithms.
