Model Replication in the Context of Agent-Based Simulation: Lessons Learnt from Two Case Studies

Abstract

This paper examines model replication in the context of agent-based simulation through two case studies. Replication of a computational model and validation of its results is an essential tool for scientific researchers, but it is rarely used by modelers. In our work we address the question of validating and verifying simulations in general, and summarize our experience in approaching different models through replication with different motivations. Two models are discussed in details. The first one is an agent-based spatial adaptation of a numerical model, while the second experiment addresses the exact replication of an existing economic model.