Multi-attribute Reverse Auction-based Protocol for Resource Allocation in Grids

Auction models are found efficient in managing resources allocation, which is a key technology in grid computing system. In this paper, a new reverse auction approach is proposed, and its mechanism and related auction algorithm are designed. In our approach, the grid trading service, i.e., an intelligent entity, uses this auction algorithm to estimate the resource user’s satisfaction degree and help the resource providers to update their multi-attribute bidding strategies in the next round. Numerical simulating experiments show that our model can satisfy a resource user’s quality demand on multiple attributes, and have better performance in user utility. The results also illustrate that revealing the resource user's preferences increases allocated efficiency.