A BBO based Procedure for Evolving Fuzzy Rules of Relaxed-criteria Negotiation in Grid Resource Allocation

Designing the negotiation agents and equipped them with the fuzzy decision controller to determine the relaxation amount in the face of intense grid market pressure leads to enhance both success rate and speed of negotiation. However, the marketoriented grids are unpredictable as new opportunities and threats are constantly being introduced as grid resource consumers and owners enter and leave a market. According to the grid market dynamics, it is needed to design adapting and self organizing negotiation agents that not only have the flexibility of relaxing the bargaining criteria using fuzzy rules but also have the ability to evolve their structures by learning and adapting new relaxedcriteria fuzzy rules. The impetus of this work is designing new negotiation agents in name Ev_MBDNAs that have two distinguishing features: 1) relaxing their bargaining term using Fuzzy Grid Market Pressure Determination System and 2) evolving their structures by learning new relaxed-criteria fuzzy rules to enhance their negotiation performance as they participate in a series of e_markets. The second feature of Ev_MBDNAs is provided by designing an evolutionary procedure that invokes Biogeography-based optimization (BBO) algorithm. In our experiments, we compare the proposed Ev_MBDNAs with EMBDNAs (i.e., negotiation agents with fixed relaxed-criteria fuzzy rules). The results show that by designing a BBO-based evolutionary procedure for learning effective relaxed-criteria fuzzy rules, Ev_MBDNAs generally outperformed EMBDNAs in different types of e_markets.