Abstract
Agent-based simulation is one of the potential approaches for understanding complex organizational and social phenomena. However, this approach has a serious problem characterized as a lack of cumulative progress. To tackle this problem, this paper investigates the main causes of the problem and explores some solutions. Through an analysis, the following implications are revealed: (1) cumulative progress in an agent-based approach is promoted by (a) the sharing of common results, (b) the development of standard computational models, (c) the replicating of older works, and/or (d) the creation of standard evaluation criteria; (2) our approach has great potential for promoting cumulative progress in an agent-based approach in terms of achieving the above four points; and (3) the factors found in our approach have the high possibility of being a fundamental process in the real world, supporting the KISS principle, and being utilized as toolkits.
