现有的基于信任的推荐模型中,交互请求者都会选择当前信任度最高的对象进行交互,这会造成虚拟社区中的资源不能被充分利用,可信度高的对象也会因为超负荷而造成响应延时,引起信任度的下降.为了解决该问题,提出了一个考虑社区节点信任度稳定性与社区利益的推荐算法.该算法采用李雅普诺夫优化的思想,社区控制中心根据目前各成员节点的状态与请求者的特征做出满足请求者的约束条件,同时保证社区节点信任度稳定性,并使得社区所得利益最大的决策.对算法进行理论分析与仿真实验表明,与传统方法相比,该算法使得社区节点的信任度与工作负荷更加稳定,同时可使整个虚拟社区与资源请求者获得更多的利益.
In available trust based recommendation schemes, users always tend to transact with the entity with the highest trust value, which will result in the fact that the virtual community's resource cannot be fully utilized. The entity with the highest trust value will also delay the response because of its capacity constraint, so that its trust value will be bound to decline. To resolve this problem, we propose a recommendation algorithm which takes the community benefit and the stability of the entities' trust value into account. The community's control center adopts the Lyapunov optimization method to make a decision which can give it the maximum benefit on the premise of satisfying the desired constraint of the interaction requester, meanwhile guaranteeing the stability of the community nodes' trust value. Theoretical analysis and experiments show that our algorithm can stabilize the community nodes' trust value, and that it also can enable the whole community and resource requester to get more benefits.