| Computation Market (CM) is an auctioning based resource trading/acquiring system that can be deployed in wide-area computing systems such as Grid systems. This study mainly focuses on the auction market design and experiments. The auction market design entails designing an economic model, i.e., protocols, rules, policies, and strategies, by using game theory. The CM adopts a multi-unit computational auction (MUCA) which allows the users to bid on multiple items at one time and then allocate in an “all or none” manner. The MUCA offers the CM the capability to allow the users to bid flexibly, efficiently, and easily. However, it also brings a problem of computational complexity-determining the winners for a MUCA problem is NP-hard. In this thesis, we propose a set of heuristics in conjunction with defining rules and policies to solve the MUCA problem. Simulations are performed and results are analyzed to evaluate the heuristics. |
