Double Learning Algorithm in Bilateral Multi-Issue Negotiation under Incomplete Information

In order to enhance the efficiency in bilateral multi-issue negotiation under incomplete information, double learning algorithm that includes Q-learning algorithm and Bayesian learning algorithm is presented. The Q-learning algorithm is used to learn the weights of issues, and the Bayesian learning algorithm is used to learn the reservation price of issues. Experiments show that the algorithm can help agents to negotiate more efficiently.

Classification-based self-adaptive differential evolution and its application in multi-lateral multi-issue negotiation

Multi-lateral multi-issue negotiations are the most complex realistic negotiation problems. Automated ap- proaches have proven particularly promising for complex ne- gotiations and previous research indicates evolutionary com- putation could be useful for such complex systems. To im- prove the efficiency of realistic multi-lateral multi-issue ne- gotiations and avoid the requirement of complete informa- tion about negotiators, a novel negotiation model based on art improved evolutionary algorithm p-ADE is proposed. The new model includes a new multi-agent negotiation protocol and strategy which utilize p-ADE to improve the negotia- tion efficiency by generating more acceptable solutions with stronger suitability for all the participants. Where p-ADE is improved based on the well-known differential evolution （DE）, in which a new classification-based mutation strategy DE/rand-to-best/pbest as well as a dynamic self-adaptive pa- rameter setting strategy are proposed. Experimental results confirm the superiority of p-ADE over several state-of-the-art evolutionary optimizers. In addition, the p-ADE based multi- agent negotiation model shows good performance in solving realistic multi-lateral multi-issue negotiations.

Automatic Identification of Customized Instruction Based on Multiple Attribute Decision-Making for Multi-Issue Architectures

This paper illustrates the importance of the configuration of function units and the change of an application’s critical path when using instruction set extension （ISE） with multi-issue architectures. This paper also presents an automatic identification approach for customized instruction without input/output number constraints for multi-issue architectures. The approach identifies customized instructions using multiple attribute decision-making based on the analysis of several attributes for each candidate node. Tests indicate that the approach achieves higher speedup ratios than previous approaches, as well as less area cost. In addition, this approach provides designers with multiple candidate designs.

EFFECTIVE AUTOMATED NEGOTIATION BASED ON ISSUE DENDROGRAMS AND PARTIAL AGREEMENTS

Negotiation is both an important topic in multi-agent systems research and an important aspect of daily life. Many real-world negotiations are complex and involve multiple interdependent issues, therefore, there has been increasing interest in such negotiations. Existing nonlinear automated negotiation protocols have difficulty in finding solutions when the number of issues and agents is large. In automated negotiations covering multiple independent issues, it is useful to separate out the issues and reach separate agreements on each in turn. In this paper, we propose an effective approach to automated negotiations based on recursive partitioning using an issue dendrogram. A mediator first finds partial agreements in each sub-space based on bids from the agents, then combines them to produce the final agreement. When it cannot find a solution, our proposed method recursively decomposes the negotiation sub-problems using an issue dendrogram. In addition, it can improve the quality of agreements by considering previously-found partial consensuses. We also demonstrate experimentally that our protocol generates higher-optimality outcomes with greater scalability than previous methods.