Knowledge-Based Simulation at the RAND Corporation

摘要: Despite its ubiquity, current discrete-state simulation technology is sharply limited in power, applicability, and credibility. Simulation models are typically to answering questions of the form “What if... ?” cannot answer other kinds at least equal value. The lack explicit knowledge most simulations makes it difficult verify correctness their underlying them hard comprehend. Recent advances object-oriented languages, rule-oriented approaches, logic programming, automated inferencing, interactive graphics facilitating a new generation “knowledge-based” environments. synergy these techniques promises revolutionize simulation, transforming into something far more powerful, useful believable. This chapter describes this emerging context Knowledge-Based project RAND Corporation, which has pursued research number related areas, including reasoning representing multiple relationships among simulated entities, highly interfaces, sensitivity analysis, varying level aggregation model, modeling “soft” concepts (such as initiative).