Auction-based models and protocols introducemoney and pricing as the technique for coordinationbetween users and producers of resources in grids. In thispaper, we propose a multi-attribute reverse auction protocolto allocate grid resource. In order to enhance theinformation efficiency we develop a novel interactiveapproach supporting both user agents and producer agents.In this approach, an iterative algorithm used to estimate theresource user’s preference utility function can facilitateproducer agents to update their bids with incompleteinformation. Numerical simulating experiments show thatour approach can satisfy the resource user’s quality demandon multiple attributes, and achieve high efficiency in userutility. The results also provide an idea to analyze the twosideprivate information problem in gird resource allocation.
展开▼