An Interactive Approach for Multi-Attribute Resource Allocation in Grids

Abstract

Auction-based models and protocols introduce money and pricing as the technique for coordination between users and producers of resources in grids. In this paper, we propose a multi-attribute reverse auction protocol to allocate grid resource. In order to enhance the information efficiency we develop a novel interactive approach supporting both user agents and producer agents. In this approach, an iterative algorithm used to estimate the resource user’s preference utility function can facilitate producer agents to update their bids with incomplete information. Numerical simulating experiments show that our approach can satisfy the resource user’s quality demand on multiple attributes, and achieve high efficiency in user utility. The results also provide an idea to analyze the two-side private information problem in gird resource allocation.