Resource trading using cognitive agents: A hybrid perspective and its simulation

In this paper, we explore a market-based grid resource trading system from a cognitive perspective. First, we introduce a six-layer framework for simulating the trading system of grid computing resources. Then we analyze a trading case: resource advance reservation through agents participating in multiple sequential auctions. We study the evolution of this trading system using cognitive agents that can automatically adapt to the environment, exchange private information and learn new experiences from their network neighborhoods. We compare the different outcomes of the experiments in non-cooperative and local cooperative perspectives. Finally, after analyzing the experiment results, some critical issues regarding the design of market-based trading system are discussed.