Analysis of the use of events and states as brute facts in modelling of institutional facts

Abstract

Although the institutional dimension of a multi-agent system can be affected directly by the actions of the agents, it can be also affected by facts originating in the environment or even in the institution. In previous work, we proposed a model, language and its interpreter to specify the institutional consequences of both events and states from environment and institution. This paper analyses this twofold approach, looking for a better understanding about the performance of the interpreter and about the design differences between using event and states. The contributions of this work are (i) the evaluation of some aspects of a proposed and implemented language, (ii) guidelines to choose between events and states to model count-as rules, and (iii) an initial benchmark to evaluate further improvements to the interpreter and the performance of similar proposals. © 2014 Springer International Publishing.