Lenient negotiation model based on altruistic utility and its implication on agent-mediated negotiation

摘要

Negotiation is a widespread business activity which is used to satisfy mutual interests to reach a consensus. With the advent of intelligent software agents, fully automated negotiation systems are achieved where negotiation is conducted between software entities without any human intervention. Of the significant challenge faced by developers is the construction of agent's proficiency to imitate human negotiators. There are a wide variety of negotiation approaches and tactics that have been proposed such as game theory, evolutionary, fuzzy and heuristic but they mostly rely on the utilitarian decision model which mainly target to maximize selfish-pleasure. Consequently, business bargains nowadays become more oriented to capital and revenue and very little attention is given to human ideals and ethics. To a certain extent this reflects the dominance of secularism in the world. In addition to contradicting the moral injunctions that Islam and most religions encourages, the assumptions behind self-maximization lacks the empirical prove. In contrast, altruism proved its persistence in between the participants engaged in many experiments. Inspired by the concept of Islamic leniency this paper propose a negotiation model that rely on altruistic utility instead of the dominant selfish utility. The proposed model is implemented through a web-based e-marketplace and then evaluated through a number of experimental tests. The results show the influence of altruism perception on the negotiation outcome and hence affect agent's behavior toward offering more affordable offers.