Community Research and Development Information Service - CORDIS

Novel interaction solutions for multi-agent systems

In multi-agent systems (MASs), humans and software interact to solve problems that are beyond the individual capacities or knowledge of each problem solver. An EU initiative examined the incompatibility of goals between individual agents that obstruct problem solving processes and lead to ineffective outcomes.
Novel interaction solutions for multi-agent systems
MASs are made up of several agents that exist at the same time, share common resources and communicate with each other. MASs attempt to formalise the coordination between agents and to understand how independent procedures can be organised. Conflicting goals among individual agents such as computer programs, robots or humans can deter or encourage the achievement of end goals.

To address the issue, the EU-funded NINA (Norms in action: Designing and comparing regulatory mechanisms for multi-agent systems) project set out to provide solutions on how to regulate and resolve potential conflict in MASs by studying computational methods.

Project partners modelled conflict and associated practices that are needed for successful resolution. Specifically, they devised models of norms as endogenous contracts and as exogenous mechanisms. This was followed by an algorithmic analysis of both norm models.

To model conflict and goal-directed behaviour, the NINA team identified and analysed a class of logic-based Boolean games. It also identified an established model of pre-play negotiation found in endogenous games.

The framework of Boolean games was enhanced with side payments that originate from game theory. Researchers analysed equilibria in the two game theoretical models and demonstrated the conditions required to achieve anticipated outcomes without any outside intervention. They used taxation/negotiation mechanisms to show how to transform endogenous Boolean games in order to obtain desirable results without the need for side payments.

NINA introduced an effective resolution procedure that proposes what best suits each interaction between individual agents. Results translate to all facets of MASs, from algorithms to game theory and beyond.

Related information


Multi-agent systems, norms, regulatory mechanisms, Boolean games, endogenous games
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top