A significant amount of research in the field of multi-agent systems is currently focused on the theory, design and implementation of 'electronic institutions'. This work adapts the mechanisms that keep human society orderly to provide a framework for building open systems of self-interested software agents that are subject to explicitly defined rules of behaviour. Three key requirements in this area are languages for expressing the norms or expectations that apply to agents' interactions and actions, techniques for detecting violations of these rules of society, and mechanisms to prevent or discourage such violations.
A number of e-institution specification languages have been proposed in the literature, but these have a limited ability to express social expectations with complex temporal constraints. This talk will describe a more expressive language for defining conditional social expectations based on a temporal logic, and will show how activated expectations can be monitored at run time to check for their fulfilment and violation.
Last modified: Thursday, 28-Jul-2005 17:23:30 NZST
This page is maintained by the seminar list administrator.