Iterative Combinatorial Auction for Two-Sided Grid Markets: Multiple users and Multiple Providers
Heterogeneity and different ownerships of grid computing resources impose complexity in evaluating the market value of these resources. Auction protocols are proposed to meet this challenge efficiently. Auction models are also suitable for achieving better payoff and resource allocation for grid providers. Grid users and providers are usually geographically distributed. The number of users in grid computing could also be very high. Hence, models provide seamless support to multiple users and providers would be useful to promote grid computing. In this paper, we implement a novel First Price Open Cry auction (ascending-bid auction) that supports for multiple users and providers simultaneously. We explain about (i) bundle generation (resource packages by providers), (ii) creating corresponding agents to bundles, (iii) allowing users to choose their suitable bundles, and (iv) clearing bundles through solving winner determination problem. The simulation results predict when and how to map providers’ private values on resource bundles, such that maximum revenue and better utilization of idle resources. K
