Context Aware Win-Win Situation For Automatic Negotiation In Intelligent Agents

Abstract

Negotiation is an important human activity in almost every matter of daily life to reach an agreement. Several automated negotiation support systems have been developed to help ease the negotiation process in human-to-human or human-to-agent negotiation. The major drawback of available systems is that they mainly focus on an agent to agent communication, which may not be suitable in an environment where humans make decisions, particularly in multi-literal negotiations. This research paper aims to present a Context-Aware Win-Win Situation for Automatic Negotiation (CAWSAN) for mutual negotiation between human and intelligent agents. Proposed system will be used for automated negotiations using a win-win strategy where preference elicitation will be performed by human negotiator while offer generation and offer evaluation will be handled by autonomous agent. In negotiation, a win-win strategy is a kind of compromise/cooperation where any negotiator generates an offer which is more acceptable for his/her counterpart by compromising claims on one matter to get extra benefit on other issues. A heuristic computational model is proposed in this research to simulate win-win strategy. The proposed computational model will employ a novel linear algorithm to estimate the preferences of other participants by using hill climbing approach to search the space of possible tradeoffs that are more acceptable by other negotiators.