Hunan University/College of Information Science and Engineering, Changsha, China

---

Abstract—In multi-agent systems, agents can cooperate with each other in order to form groups and take advantage of lower prices without ordering more than their actual demand. In this paper, we study a cooperative search of autonomous agents that represent agents’ coalition formation to enjoy a price discount for each of its requested service to achieve a goal. Such cooperation among agents is not only with direct agent’s neighbors, but also with indirect agent’s neighbors through the concept of agent graph, where several opportunities can be found, and each opportunity is associated with a value for the coalition. Given a search cost, the goal of an agent is to find the best set of opportunities which fulfills the coalition’s demands for the requested services with the maximum overall utility. The experimental results illustrated the performance of proposed Semi-Optimal Discount Group Formation (Semi- Optimal DGF) model and have proven that, searching in groups can exploit opportunities that have been discarded by individual agent's search to reduce the percentage of failure for achieving goals to zero. At the same time, the proposed model proved its efficiency in average processing time from 5% to 24% faster when compared with other optimal models according to a different number of agents that form coalitions. With increased number of agents, the proposed model has an average utility 98% closest to the optimal utility and higher than the average coalition value obtained by 6% when considered a search cost as a parameter affected on the search for optimal coalitions.

Index Terms—Coalition Formation, Multi-Agent System, Cooperative Search, Agent Graph

[PDF]

Cite: Walaa H. El-Ashmawi, Jun Hu, and Renfa Li, " Proposed Discount Group Formation Model based on Cooperative Search in Agent Graph," Journal of Computers vol. 8, no. 10, pp. 2497-2506, 2013.