Verification of Artifact-Centric Multi-Agent Systems

Title: Verification of Artifact-Centric Multi-Agent Systems
Speaker: Alessio Lomuscio (Imperial College London, UK)
Time: 10:30, Monday, November 12th, 2012
Venue: Lecture Room, 3rd Floor, Building #5, State Key Laboratory of Computer Science, Institute of Software, Chinese Academy of Sciences
Abstract:
Artifact-Centric systems are a particular kind of web-services where data feature prominently in the system description. The emphasis on data makes automata-based formalisms commonly used to model services insufficient and calls for the explicit representation of the evolution of the underlying databases. In this talk I will explore the verification problem for artifact-centric multi-agent systems, i.e., systems of agents interacting through artifact systems. I will point to the undecidability of the model checking problem of these systems when analysed against specifications based on first-order temporal-epistemic logic. I will then analyse conditions that enable us to obtain a decidable problem through finite abstractions that are bisimilar to a given model. The talk is based on results published at IJCAI2011, ICSOC2011, and KR2012.

Bio sketch:
Alessio Lomuscio is a Professor in logic for multi-agent systems in the Department of Computing, Imperial College London, where he leads the VAS (verification of autonomous systems) research group. He currently holds an EPSRC Leadership Fellowship. He received a PhD in Computer Science from the University of Birmingham in 1999 and a Laurea in Electronic Engineering from Politecnico di Milano in 1995. Before joining Imperial he was Lecturer at King’s College London and Senior Lecturer at University College London. His research interests concern the specification and verification of multi-agent systems by means of techniques based on computational logic. In particular he has made theoretical contributions in the area of logic for multi-agent systems (including studying the completeness, decidability and complexity of several temporal-epistemic formalisms) and has put forward several symbolic model checking techniques for these.  He has applied these techniques to the verification of a range of applications including autonomous underwater vehicles, web-services and security protocols including e-voting.  He is co-author of MCMAS, an open-source, BDD-based model checker for multi-agent systems developed at the VAS research group at Imperial.