Abstract: Lack of reliable trust evaluation mechanism prevents resources from the comprehensive sharing in grid computing economy.Based on grid computing economy model，a trust model with controlled evaluation is presented，in which trust context is introduced and corresponding incentive and punishment mechanism is put forward.With the estimation of context similarity，the model effectively estimates trust degrees of strange entities.Simulation shows that the model outperforms the existing trust evaluation model，and can well solve the reliable trust evaluation problem of grid entity，decrease the influence of strange entities’ pseudo services on trust evaluation models.

Key words: grid computing economy, trust, context, Euclid approximation, incentive

摘要： 当前的网格计算经济缺乏可靠的信任评价机制，这会影响网格资源的共享。在网格计算经济模型基础上，提出了信任评价控制模型。模型引入信任上下文的概念，通过上下文相似度估计，有效估算陌生实体的信任度，依据实际交易情况对交易实体提出激励处罚机制。仿真结果表明，提出的模型优于现有的信任评价模型，能够较好地解决网格实体信任评价可靠性问题，降低了陌生实体的虚假服务对信任评价模型的影响。

关键词: 网格计算经济, 信任, 上下文, 欧几里德近似法, 激励