一种改进的粒子群优化算法及其在盲信号分离中的应用

首先对粒子群优化算法进行了改进，用粒子群中所有粒子最优位置的加权平均值替代速度更新式中的粒子最优位置，并根据粒子群中粒子的浓度自适应地调整加速系数．这种改进的粒子群优化算法利用了粒子群中所有粒子最优位置的信息，并在进化过程中自适应地调整“认知”部分和“社会”部分对粒子的影响，从而提高了算法的收敛速度和精度．然后，针对盲...

An improved particle swarm optimizer and its application in blind signals separation

In this paper, an improved particle swarm optimizer is first presented by replacing personal best particle with the weight average of personal best particles in swarm, and adjusting adaptively acceleration coefficients based on particle＇ s density in swarm. The proposed algorithm applies more particles＇ information, and adjusts adaptively ＂cognition＂ component and ＂social＂ component, thus improves convergence performance. Then, based on second order correlation coefficient and high order correlation coefficient, a cost function of independence of signals is proposed, and an algorithm for linear blind signals separation is presented by applying the improved particle swarm optimizer to the cost function. The experiment results demonstrate that the improved particle swarm optimizer is superior to original particle swarm optimization, and the algorithm for linear blind signals separation is efficient.