Using model replication to improve the reliability of agent-based models

  • Authors:
  • Wei Zhong;Yushim Kim

  • Affiliations:
  • Arizona State University, Phoenix, AZ;Arizona State University, Phoenix, AZ

  • Venue:
  • SBP'10 Proceedings of the Third international conference on Social Computing, Behavioral Modeling, and Prediction
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

The basic presupposition of model replication activities for a computational model such as an agent-based model (ABM) is that, as a robust and reliable tool, it must be replicable in other computing settings. This assumption has recently gained attention in the community of artificial society and simulation due to the challenges of model verification and validation. Illustrating the replication of an ABM representing fraudulent behavior in a public service delivery system originally developed in the Java-based MASON toolkit for NetLogo by a different author, this paper exemplifies how model replication exercises provide unique opportunities for model verification and validation process. At the same time, it helps accumulate best practices and patterns of model replication and contributes to the agenda of developing a standard methodological protocol for agent-based social simulation.