摘要： 为解决时间信号盲源分离算法中的时延选择问题，提出一种基于量子遗传算法的时延自适应优化选择方法。采用量子编码表征染色体，量子坍塌的随机观察结果与时延相结合形成种群，对若干时延二阶相关矩阵同时近似对角化，利用分离信号的负熵构造适应度函数，通过量子旋转门算子来实现染色体的演化更新。语音信号的盲源分离实验结果表明，与其他方法相比，该方法具有更好的种群多样性和更快的收敛速度及全局寻优的能力。
关键词:
量子遗传算法,
量子计算,
盲源分离,
TDSEP算法,
时间信号,
时延优化
Abstract: Aiming at the optimal selection problem of time delay for temporal blind source separation, a novel method of optimal selection of time lags based on Quantum Genetic Algorithm(QGA) is proposed. The chromosome is adopted. The random observation the quantum collapse combines with time delay to structure populations. After simultaneous approximate diagonalization of several second order correlation matrices, the fitness function based negentropy is constructed. Chromosome update is achieved through quantum rotating gate. Simulation experimental result shows that QGA has more rapid convergence and better global search capacity than other algorithms.
Key words:
Quantum Genetic Algorithm(QGA),
quantum computation,
blind source separation,
TDSEP algorithm,
time signal,
time delay optimization
中图分类号:
曹军宏, 韦灼彬, 高屹, 张宁. 基于量子遗传算法的盲源分离时延优选[J]. 计算机工程, 2012, 38(11): 170-172,176.
CAO Jun-Hong, HUI Zhuo-Ban, GAO Ge, ZHANG Ning. Optimal Selection of Time Delay for Blind Source Separation Based on Quantum Genetic Algorithm[J]. Computer Engineering, 2012, 38(11): 170-172,176.