Norm regulation in collaborative virtual environments by normative multi-agent systems
MetadataShow full item record
Different types of research have been done on multi-agent systems regarding normative systems. This research addresses the enforcement of norms by a multi-agent system. More specifically this thesis investigates the question whether or not it is possible for a normative multi-agent system to regulate the norms in a collaborative virtual environment. The collaborative virtual environment used in this research is called Club Time Machine and is inhabited by users with the age between 6 and 12. To ensure a good interaction between the human users and the multi-agent system, the reasoning of the system will be done based on a BDI-model which also incorporates normative rules. This BDI-model is designed and simulated with the LEADSTO software. The traces generated by the simulations of the LEADSTO software are then checked with the help of Temporal Trace Language (TTL) properties to test if the model functions properly. A framework is designed for the multi-agent system to integrate this BDI-model and let the multi-agent system inhabit the collaborative virtual environment. To test this framework and thereby the research goal a number of test cases have been defined. The framework then outputs the results of these test cases into LEADSTO traces, which can be checked with the defined TTL properties. The experiments show that the normative multi-agent system first increases the number of norm violations. However, over a period of time the users become accustomed to the agent and then they comply to the norms of the virtual environment.