Research On E-commerce Negotiation Based On Answer Set Program

With the rapid development of information technology, electronic commerce as a kind of fast and convenient way to trade, is more and more popular with people’s lives, and negotiation as an important part in e-commerce transactions, has been played a vital role in the study of multi-agent system. Negotiation research based on logic program currently is a research focus in the multi-agent negotiation, however, answer set program as a special logic program, has a wealth of knowledge expression and non-monotonic reasoning ability. This paper will use answer set program in the automatic negotiation of e-commerce study, and put forward an asynchronous negotiation model, in which only one party of the negotiation participants(the buyer) modified its answer set program. In this model, answer set program as the representation tool that represents agents ’ knowledge, and each agent choose their a answer set of answer set program as their initial negotiation requirements, then, through the consumer give up their part needs and receive the seller’s requirements to simulate the process of the whole negotiation process. As a party of the negotiation, in order to ensure the transaction efficiency and avoid the logic program updating frequently, electric business platform(the seller) do not change their answer set program in the whole course of the negotiation, but it will update its answer set program after a period of time, and generate needed updating program through past consumer demand information In the transaction records, then merge the updating program to the buyer’s program to complete the update.For every rational agent, when they choose an answer as the initial negotiation requirements, they will have their own preferences, these preferences are usually the key to decide whether the negotiation will success or not, because these decide which requirements can be given up, and which requirements they will not compromise even if the negotiation fail. At the time of initial consultation, both parties of negotiation will want to rank their negotiation requirements according to their own preferences, and select an answer set which most can reflect their preferences as their initial consultation requirements. In this paper, by analyzing the dependency relation between the rules of answer set program, we construct a rules dependency graph with these dependent relations. and put forward a negotiation requirements sorting method based on rule dependency graph. And requirements of agent which the put forward as they initial negotiation are classified by core requirements and non-core requirements, for agent, non-core requirements are able to be given up, and core requirements can’t be given up.As the process of negotiation is the process of game, through the analysis of the game theory properties of negotiation model, verifying the result of the negotiation whether has some equilibrium properties of the game theory is a valid way to estimatethis negotiation model effective. Therefore, in this paper, we demonstrates that the proposed asynchronous negotiation model has the following several relevant game theory properties:In the condition that the seller updating his answer set program in the process of trading will increase the cost, no matter if there exists conflicts, the nash equilibrium results of the game between the seller and the buyer are: the buyer purchase goods, and the seller don’t update his answer set program, so this proves the correctness of the asynchronous negotiation mode which put forward in this paper.With this asynchronous negotiation model, the game between the seller and the buyer is perfect information dynamic game and there exists a sub-game refining nash equilibrium solution.Finally, we give an e-commerce negotiation system prototype based on the answer set program, analyze the process of negotiation between the seller and the buyer and the update method of the seller’s answer set program.