Research On Bilateral Simultaneous Automated Negotiations For E-Marketplaces

Traditional face-to-face negotiation is the most predominant way for solving bilateral negotiation problems, and has a certain flexibility. However, humans, due to limited information-processing capabilities and biases, seldom perform effectively during negotiations when the problems involved are intractable. With rapid growth of the Internet, trading takes place on electronically with the help of the automated agent has been more and more widespread. For now, there have emerged many research works about automated negotiation. The alternating-offer bargaining protocol is adopted in the most of works. However, there are also some disadvantages using this method. Typically, a deadlock may asire in which the negotiators are not willing first to bid or concede. There are some research about simultaneously offer through a mediation. But in these studies, both agents need to reveal their preferences to the Mediator agent which doesn't meet the need of actual electronic marketplace. It is more and more urgent to design an automated negotiation system meeting the need of electronic marketplace. On the basis of a full review of the state of bilateral negotiation theory and application, this paper is to probe into the bilateral simultanous autoumated negotiation problems using the game theory with actual electronic marketplace in the background.Bilateral single-issue negotiation is the basic of all negotiation problems. In this paper, a bilateral simultanous automated single-issue negotiation mechnism is proposed. Negotiation mechnism is composed of negotiation protocol and negotiation strategries. In the single-issue bargaining protocol, both agents simultaneously submit their respective offers to the mediator agent by introducing a nonbiased mediator. In our agent-based model, to avoid strategic misrepresentations, the offer is not revealed to the opponent but only to the Mediator agent. The utility function is a key factor for influencing the strategy. In the negotiation model, we establish the utility function for both agents first, and then analyze how an agent can exploit the available information to select a strategy that maximizes its expected utility and give the optimal offer of each agent at each stage of the negotiation. This mechnism avoids the insufficiency of the traditional alternating-offer pattern and can discourage counter-speculation and effectively control fraud and misrepresentation in a certain extent.Based on the proposed single-issue negotiation mechnism, a bilateral simultanous automated multi-issue negotiation mechnism is proposed. Similarly, in the multi-issue bargaining protocol, both agents simultaneously offer by introducing a nonbiased mediator. In this paper, all negotiated issues are divided into two categories:continous issues and discrete issues. The proposed model allows concurrent negotiation threads for accelerating the negotiation process. A combination of all discrete issues corresponds to a negotiation thread. In the context of each thread, both agents can only respond by varying the offers for the continous issues. The thread in which the agreement is reached first determines the final agreed result. Then a mathematical model is established by considering an individual dealer (Buyer) stocking to a supplier (Seller). In this model, both agents are not conflicting over the quantity issue, i.e.,a Seller may desire a larger quantity but a Buyer doesn't necessarily desire a smaller quantity. After establishing the utility function for both agents, we analyze how an agent can exploit the available information to select a strategy that maximizes its expected utility and give the optimal offer of each agent at each stage of the negotiation. Our proposed mechnism avoids the insufficiency of the traditional alternating-offer pattern, and both agents need not to reveal their preferences to the Mediator agent which meets the need of actual electronic marketplace.In order to test the feasibility and effectiveness of the proposed negotiation mechnisms, an experimental study is carried out by using MATLAB tool.The experimental results demonstrate that the proposed automated negotiation mechnisms are efficient in terms of several general evaluation criteria, which is desired for the negotiators.The proposed negotiation mechnisms avoids the insufficiency of the traditional alternating-offer pattern, and consider two situations containing bilateral single-issue and multi-issue, covering a wide range. And the experimental results demonstrate the negotiation mechnisms are efficient, but it is a good negotiation method in actual electronic marketplace.