CAJ | 학술논문

　　简要介绍了果蝇优化算法的基本理论，针对FastICA等算法的稳定性和收敛性不够，而粒子群优化的盲分离运算速度慢的问题，将改进的果蝇优化算法应用到盲源分离研究中，提出了一种基于改进的果蝇优化的盲源分离算法。算法以信号的规范四阶累积量为代价函数，以改进果蝇算法对代价函数求极值，逐一确定分离向量，完成对线性瞬时混合语音信号的分离。仿真结果表明，算法能够有效实现对各混合语音信号的有序盲分离，且分离顺序能够确保按照源信号的规范四阶累积量绝对值的降序进行，分离精度也有一定的提高。
　　간요개소료과승우화산법적기본이론，침대FastICA등산법적은정성화수렴성불구，이입자군우화적맹분리운산속도만적문제，장개진적과승우화산법응용도맹원분리연구중，제출료일충기우개진적과승우화적맹원분리산법。산법이신호적규범사계루적량위대개함수，이개진과승산법대대개함수구겁치，축일학정분리향량，완성대선성순시혼합어음신호적분리。방진결과표명，산법능구유효실현대각혼합어음신호적유서맹분리，차분리순서능구학보안조원신호적규범사계루적량절대치적강서진행，분리정도야유일정적제고。
The basic theory of Fruit Fly Optimization Algorithm(FOA)is introduced. According to the stability and Conver-gence of FastICA is not enough, and the Blind Source Separation(BSS)of particle swarm optimization is not fast enough, the improved fruit fly optimization algorithm is being applied in researching of BSS, and a new BSS algorithm is proposed. The ab-solute value of normalized fourth-order cumulant is used in the algorithm as cost function and the improved Fruit Fly Optimiza-tion Algorithm is used to optimize the cost function. One by one to determine the separation vector and the instantaneous linear mixed blind source audio signal could be separated in sequence. In the meantime, the simulation results show that the algorithm can achieve the efficient sequential blind separation for source signal and ensure the separation order according to the descend-ing absolute value of normalized fourth-order cumulant, and has higher separation accuracy.