Partner Selection for Agents: A Utility Theory Approach

Abstract

This paper proposes an idea for modeling of matching between agents based on utility theory. The basic idea comes from the stable matching problem, and this focuses on how an agent makes an ultimate decision when it comes to finding a matching agent based on utilities that other agents have as well as the agent's preferences for each utility option. This paper begins with introductions of agent organization, partner selection in Agent- Based Model, and utility theories. The proposed approach of this paper includes aspects of directions and applications. In terms of the aspect of directions, pre-defined concepts such as utility options, perceptual and actual utility, and deal-breakers are introduced to understand applications better. Also, two possible scenarios with basic and high complexity were handled as applications for partner matching between agents based on utilities. There are three merits of the proposed approach, which are: (1) both the agent's benefit and its potential partner agent's benefits are considered for mate selection; (2) by employing Congregation Organizational Paradigm, an agent's partner selection approach can be applied in dynamic environments by seeking its optimal partner in another networking cluster; and (3) by employing threshold and deal-breakers, a rational agent can avoid making a decision which lowers its level of happiness or benefits.