After discussing a potential research project with a colleague this week, I was reminded of two papers from my (pre-blogging) PhD days, applying logic programming to multi-agent systems. This work used answer set programming (ASP) — a form of declarative programming oriented towards difficult (primarily NP-hard) search problems, based on the stable model semantics of logic programming — to reason about games in multi-agent systems. These two papers with colleagues from Bath present working examples of how ASP can be used for real-world reasoning, allowing agents to individually reason and communicate their knowledge and beliefs in a range of scenarios and domains.
The first paper was presented at the 3rd International Workshop on Declarative Agent Languages and Technologies (DALT III) in Utrecht in July 2005 and published in LNCS 3904:
LAIMA: A Multi-Agent Platform using Ordered Choice Logic Programming
Marina De Vos, Tom Crick, Julian Padget, Martin Brain, Owen Cliffe and Jonathan Needham
Multi-agent systems (MAS) are able to take many forms depending on the characteristics of the agents populating them. Amongst the more demanding properties from a design and implementation point of view is how these agents may individually reason and communicate about their knowledge and beliefs, with a view to cooperation and collaboration. In this paper we present a deductive reasoning formalism based on a multi-agent extension of answer set programming (ASP). We show that it is capable of dealing with, on a theoretical level, the specification and implementation of the system’s architecture, communication and the individual reasoning capacities of the agents. Agents are represented as an Ordered Choice Logic Program (OCLP) for modelling their knowledge and reasoning capacities. Communication between the agents is regulated by uni-directional channels transporting answer sets. For the implementation of our system we combine the strength of the JADE framework with the efficiency of the OCT answer set solver. The power of this approach is demonstrated by a multi-agent system reasoning about equilibria of extensive games with perfect information.
The second paper was presented at the 3rd European Workshop on Multi-Agent Systems (EUMAS 2005) in Brussels in December 2005 (see full proceedings):
T-LAIMA: Answer Set Programming for Modelling Agents with Trust
Marina De Vos, Owen Cliffe, Richard Watson, Tom Crick, Julian Padget, Jonathan Needham and Martin Brain
In a time where multi-agent systems (MAS) become increasingly more popular, they come in many forms and shapes depending on the requirements of the agents that need to populate them. Amongst the more demanding properties with respect to the design and implementation is how these agents may individually reason and communicate about their knowledge and beliefs, with a view to cooperation and collaboration. With information coming from various sources, it becomes vital for agents to have an idea how reliable these information providers are, especially if they start to contradict each other. In this paper we present a hybrid multi-agent platform, called T-LAIMA, using an extension of answer set programming (ASP). We show that our framework is capable of dealing with the specification and implementation of the system’s architecture, communication and the individual reasoning capacities of the agents. We discuss both the theoretical framework which models a single, fixed encounter between a number of agents and the implementation that sets ups these encounters in a open multi-agent domain.
Full text link: http://opus.bath.ac.uk/5411/
(also see: Publications)