A distribution estimation algorithm for connection function based on Kendall ' sτ solution to SaaS deployment problem in cloud computing

A distribution estimation algorithm for connection function based on Kendall ' sτ solution to SaaS deployment problem in cloud computing

Gao; Huang Delong; Yau Xiangwei; Huang

The performance of the distribution estimation algorithm is highly dependent on how to estimate and sample the probability distribution, and the connection function (Copula) is a powerful tool for constructing the probabilistic distribution model. This paper presents a distribution estimation algorithm of normal connection function based on Kendall ' sτ, and the algorithm estimates Kendall ' s Tau and the relationship between the Kendall ' sτ and the correlation matrix first, the correlation matrix in the normal join function is estimated, and the joint distribution is estimated. Then, the Cholesky decomposition algorithm is used to generate new individuals for the matrix. Because of the simplicity of the normal join function, The algorithm has the advantages of simplicity and clarity. The algorithm is applied to some test functions and the SaaS deployment problem in cloud computing, the experimental results show the effectiveness of the algorithm.

A distribution estimation algorithm for connection function based on Kendall ' sτ solution to SaaS deployment problem in cloud computing