In this paper, we explore the market-based grid resource trading system from the social perspective and the collaborative computing perspective. We firstly introduce a novel framework to simulate the trading system of the grid computing resources, then we give a case study based on this framework. In the case study, an evolutionary dynamics of the resource reservation game is played by agents on the network. We study the evolution of the system with cognitive agents: Study the evolution dynamics of system with agents can automatic adapt to their environment, they can exchange private information and learn experiences from each other. Finally, after analyzing the experiment results, some critical issues about designing the market-based trading system are discussed.