Towards a Behavioural Agent-based Assistant for e-Negotiations♦

Software agents typically negotiate on behalf of their owners. For this to be effective the agent must be able to adequately represent the owners’ preferences. However, the process by which this knowledge is acquired is normally not taken into consideration. The idea of the proposed method is that a user uses an e-negotiation system and fills in the Thomas-Kilmann questionnaire which is designed to measure the conflict mode. The assisting software agent then uses an implemented model to construct the user’s concession graph and suggests the concession for each offer package e.g. in a multi-attribute negotiation. The modeling of the concession graph was designed by comparing negotiation history graphs with the five conflict styles (competing, collaborating, compromising, avoiding and accommodating). Certain patterns were found which allow to model and map the conflict styles to a concession graph. It was found that people who had high compromising styles produced a concave concession graph, people who are highly accommodating generated a linear graph, and people who had a high value for competing and compromising had a convex concession graph. ♦ This work has been partially supported by the Natural Sciences and Engineering Research Council, Canada and the Social Sciences and Humanities Research Council, Canada.