基于改进INFO算法的新型可拓云计算机性能评估模型

计算机系统的性能评价和分级是计算机高性能发展的关键环节。为消除评估过程中不确定性导致的性能等级误判，采用可拓云理论构建计算机性能评价的理论模型；其次，为兼顾等级边界的分明性和模糊性，提出一种可自适应评估对象的最优云熵计算方法，提高了模型的合理性和准确性；最后，针对模型参数的非线性求解问题，提出改进向量加权平均算法(INFO)。该算法采用差分进化策略扰动向量来产生候选解，从而优化潜在解的初始分布均匀性；并且在搜索阶段结合t分布策略，根据概率值扩大算法的搜索范围，提高算法的全局搜索能力。经实例验证，结果表明所提模型是一种有效的计算机性能评估方法。

Extension cloud computer performance evaluation model based on improved INFO algorithm

The performance evaluation and classification of computer system is a key step in the development of computer high performance. In order to eliminate the misjudgment of performance levels due to uncertainty in the evaluation process, this paper developed theoretical model for computer performance evaluation using extended cloud theory. Secondly, to take into account the clarity and ambiguity of the stratification boundary, then this paper proposed an optimal method for calculating cloud entropy that adaptively evaluated the entropy of an object, which improved the rationality and accuracy of the model. Finally, it proposed an improved vector measurement method(INFO) for the nonlinear solution of model parameters in the presence. It employed a differential evolution strategy to perturb the vectors to generate candidate solutions to optimize the initial distribution uniformity of potential solutions. In the search phase, combined with the t-distribution strategy, the search range of the algorithm could be expanded according to the probability value to improve the global search capability of the algorithm. The results show that the proposed model is an effective method for computer performance evaluation.