中文摘要：

在特定拓扑环境下，在线社交网络推荐机制的友好程度是决定推荐优劣的关键之一，其成功与否受制于关系结构与交互结构的共演化。基于此，以关系结构推动交互结构为演化前提，以“学者”交互为特定领域的在线社交网络“学者网”为应用背景，通过E-CARGO模型建模，采用角色协同方法，针对学者网用户拓扑中的水平好友关系与垂直等级关系设计并仿真了相关好友推荐机制。经实验问卷调查，分别约67.69%与86.16%的用户接受基于拓扑层次与推荐方向的好友推荐方式，相关推荐机制自然、友好，用户评价度较高。

英文摘要：

The friendliness of an online social network’s recommendation mechanism is a key for the quality of the recommendation in a special topology. The success of recommendation depends on the coevolution of the relationship structures and interaction structures. Based on this idea, it is assumed that relationship structures promote interaction structures in an online social network such as the“scholat. com” as a background, a friend recommendation mechanism is designed and simulated with the help of the E-CARGO model and the method of role cooperation, aiming at recommending the friend relationship of within a horizontal view and the hierarchy relationship within a vertical view in a scholars’ topology. Through a questionnaire survey, there are 67. 69% of interviewees who would consider horizontal friend recommendation. What is more, 86. 16% of interviewees say that they would take the hierarchical recommendation mechanism into consideration. In a word, the pro-posed friend recommendation mechanism is natural and obtains a high degree of user evaluation.