Abstract: The value of recommendation trust is evaluated by probability average method in traditional recommendation trust model.But the method has low efficiency，and is difficult to resist the unite cheat behavior.This paper presents a searching trust path model based on ant colony algorithm，which is able to choice many better independence paths by a few circles.The algorithm is able to prevent unite cheat behavior in a certain extent.At last，some emulation experiments are offered to show that the algorithm is efficient and valuable in earthy complex network environment.

Key words: ａnt colony algorithm, ｔrust, ｔrust management model

摘要： 传统推荐信任模型中单纯采用概率平均的方法得到推荐信任值，这种方法效率较低，而且难以抵抗联合欺诈行为。基于蚁群算法给出了一个寻找信任路径的算法，该算法通过多次循环选出多条较优的独立信任路径，在一定程度上可有效防止联合欺诈行为，并通过实验证明了它的有效性，适应现实的复杂网络环境。

关键词: 蚁群算法, 信任, 信任管理模型