期刊
APPLIED INTELLIGENCE
卷 53, 期 11, 页码 13763-13800出版社
SPRINGER
DOI: 10.1007/s10489-022-03962-x
关键词
Underdetermined blind source separation; Quantum Archimedes optimization algorithm; K-means clustering algorithm; Radial basis function network
This paper proposes an underdetermined blind source separation method based on the quantum Archimedes optimization algorithm, achieving higher accuracy and more reasonable initial parameter selection through solving the objective functions and initial estimation signal setting for engineering problems.
The performance of the existing underdetermined blind source separation methods is very sensitive to the initial parameters, meanwhile, the existing setting or selection methods of initial parameters need to be improved. Consequently, an effective underdetermined blind source separation method is proposed in this paper to solve the above engineering problems. Based on the Archimedes optimization algorithm and quantum computing theory, this paper proposes a novel intelligent algorithm named quantum Archimedes optimization algorithm, which solves the objective functions for engineering problems. Then the optimal solution obtained through the quantum Archimedes optimization algorithm is used as the initial clustering centers of the K-means clustering algorithm to achieve mixing matrix estimation. In addition, the original initial estimation signal setting of the source recovery based on radial basis function network is converted into an initial solution in population for quantum Archimedes optimization algorithm. The optimal solution obtained through the quantum Archimedes optimization algorithm is used as the new initial estimation signal setting to achieve source recovery. The simulation results show that the proposed underdetermined blind source separation method has higher accuracy than previous methods. The proposed method that is more robust and applicable makes the setting and selection of initial parameters more reasonable so that the performance is no longer limited to the initial parameters.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据