摘要
为解释和修正信任管理系统中第二手信任评价与第一手信任评价之间的差异,提出一种基于DSmT的专家推荐模型——扰动模型。在该模型中,为求扰动评价,引出DSmT反问题。并介绍一种修正的粒子群优化算法,为求DSmT反问题的近似解,选取3个测试函数测试其性能,证明其能提高收敛速度。实验表明,扰动模型可以较好地解释第二手信任评价与第一手信任评价之间的差异,通过修正第二手信任评价可使其与第一手信任评价更一致。
To explain and modify the difference between the second-hand trust evaluation and the first-hand trust evaluation in trust management systems, an expert referral model, disturbing model is proposed. In this model, the inverse problem of DSmT is presented for the solution of the disturbing evaluation. In addition, a Modified Particle Swarm Optimization（MPSO） arithmetic is proposed to solve the inverse problem of DSmT, which proves to accelerate convergence greatly by a performance test with three test functions. Experimental result shows that disturbing model can explain the difference between the second-hand trust evaluation and the first-hand trust evaluation pretty well and the second-hand trust evaluation can be modified to be more accordance with the first-hand trust evaluation.
出处
《计算机工程》
CAS
CSCD
北大核心
2009年第8期217-219,共3页
Computer Engineering
基金
南京理工大学青年学者基金资助项目(NJUST200304)
