Title :
Application of Quantum Evolutionary Algorithm in Blind Source Separation
Author :
Wu, Xin-Jie ; Xu, Chao ; Cui, Chun-yang
Author_Institution :
Coll. of Phys., Liaoning Univ., Shenyang, China
Abstract :
The precision of blind source separation (BSS) by joint approximate decomposition of eigen matrices (JADE) based on fourth-order cumulants is low. In order to overcome this disadvantage, a new algorithm of BSS based on quantum evolutionary algorithm is proposed in this paper. Quantum evolutionary algorithm uses Qubit as basic information-bit for individual code, and finishes the individual evolution with unitary transformation of quantum state (quantum gate transform). At the same time, the polymorphic superposition of quantum code and whole interference crossover can overcome the prematurity in the process of evolution. Using kurtosis of the hybrid signal as the target function of BSS, the BSS method based on quantum evolutionary algorithm succeeds in separating the instantaneous hybrid signal by the method of independent component analysis. The simulation experiments have shown that the scheme is feasible and effective.
Keywords :
blind source separation; eigenvalues and eigenfunctions; evolutionary computation; matrix decomposition; quantum computing; Qubit; blind source separation; eigen matrices; fourth-order cumulants; independent component analysis; interference crossover; quantum code; quantum evolutionary algorithm; quantum gate transform; signal kurtosis; Blind source separation; Computer applications; Concurrent computing; Evolutionary computation; Independent component analysis; Matrix decomposition; Quantum computing; Quantum mechanics; Robustness; Source separation; blind source separation; kurtosis; quantum evolutionary algorithm;
Conference_Titel :
Natural Computation, 2009. ICNC '09. Fifth International Conference on
Conference_Location :
Tianjin
Print_ISBN :
978-0-7695-3736-8
DOI :
10.1109/ICNC.2009.508
