A Rationality-Based Modeling for Coalition Support

Agents Technology Research

“…Focus in these works is placed on giving the individual agent a means for selecting and keeping track of which other agents are equitable collaboration partners and which are not. It includes works that seek to determine trust values either through direct interaction only [22,33,26] and through the diffusion of reputation [30,34,29]. There are a number of open challenges associated with maintaining individual trust such as: how best to infer trust from experiences, determining defection or levels of defection from observations, modeling of opponent intentions, proper identification of agents, and the semantic exchange and interpretation of reputation.…”

“…The Affinity Management System (AMS) is our in-house developed multi-agent trust management scheme designed for cooperative multi-agent environments [26]. Agents using AMS establish trust or affinity for other agents in the system through direct interactions.…”

“…AMS determines whether or not to accept another agent's request for cooperation based on the utility to be gained from the collaboration. In [26] utility was defined as the increase in affinity, which represents the potential for reciprocity, subtracted by the cost. Cost was not fully defined and the cost of all tasks were considered equal.…”

Incentive and Trust Issues in Assured Information Sharing

“…They found the introduction of a reputation system allowed punishment systems based on discriminating between good and bad reputations succeeded as long as players were sufficiently patient to achieve the results. Other works have attempted to apply peerto-peer trust in scenarios involving military joint cooperation in the field [3]. Given a military body as a dividable resource, they attempt to address the problem of resource allocation in situations where either a central authority is not robust enough or the resources span multiple international owners.…”