| The market is a complicated system, it exists a great deal of transaction with each other, and process of transaction is frequent. So it is hard to molding with tradition methods such as analytics method, number analysis method or other formalization, semi-formalization method etc. Because all of the tradition methods of molding is through mathematics equation describes the interaction between elements, subsystems, layers and the interaction between the system and environment. The tradition methods suppose that unit of the system is homogeneous, but in the reality every unit of the system is high-heterogeneous. Particularly the market is a system which is non-linear, so if we use the tradition methods of molding will result in the mold fall away the reality. Based on the context, the paper set up the multi-Agent mold for transaction in the market, and simulate the mold in a multi-Agent simulation platform——JADE. And through the analyzing, explaining and expanding the experimental results, the paper provides the theoretic gist and guidance for the participator in the market. Following efforts are presented in this dissertation:(1) Modeling method based on the AgentThis chapter introduces provenance, concept and characteristic of the modeling method based on the Agent. And expatiating the several methods of modeling based on the agent. At the end, we expatiating the method of this paper and the advantage of the method further.(2) A study of cooperative mechanism among seller agentsThis paper proposes a cooperation mechanism among seller agents and builds a model based on multi-agent system. With this mechanism, a seller would complete the trade with the buyer although he doesn't have enough items and the stock of the seller need not too much. The model is simulated in a multi-Agent simulation platform——JADE. The experimental results show that this cooperation mechanism widens the bound of cooperation among the sellers and saves the stock cost of the sellers.(3) Simulation of Autonomous Agents for Participating in Multiple Online AuctionsThe increasing number of online auctions poses a big challenge to those who are actively looking for good deals. In this paper, we present the design of an autonomous agent that can monitors and collects information from the ongoing auctions and choose the best one to participate in. If the buyer participates in this auction he can get the maximal profit. At the end of the paper we use the JADE to simulate the whole course. Our experiments show that the success rates of the buyer can increase obviously. And the profit of the whole market is improved.(4) The research of simulation of modeling in the market with JADESimulating the two models which were introduced above in a multi-Agent simulation platform——JADE, and expatiating the process of implementing in the JADE. |
