Descriptive Agent-based Modeling of Zero-sum Games Using Cognitive Agent-based Computing Framework (T-0717) (MFN 5083)

Abstract

In complex software, modeling to understand component interactions can be a complex task. In literature, a mathematical toolkit called Game theory is used for modeling the complex minteraction of components. Game theory has a class of games called Zero-sum games. These games have been used in solving computer science problems. Zero-sum game theoretic models have been analyzed or modeled in Economics and other domains, while these games have not been analyzed or modeled from the Complex Adaptive Systems (CAS) perspective. The di erent perspectives in which these games have been used are Ecology, Biology, Economics, Politics etc. To the best of our knowledge, these games have not previously been modeled by using Cognitive Agent-Based Computing (CABC) Framework. In this thesis, Cognitive Agent-Based Computing Framework is used to develop a Descriptive Agent-Based Modeling (DREAM) for Zero-sum games. The model is also described using a formal speci- cation technique using Z. Then, we compared DREAM speci cation with Overview, Design Concept, and Details (ODD) Protocol, a textual speci cation technique. Finally, the results showed that with the DREAM speci cation of Zero-sum games, the models can be visualized, compared with other ABMs and can be translated to the source code. These things are not possible with the textual speci cation, ODD.