论文部分内容阅读
在粒子群算法的基础上引入第3方随机项:随机数和全局最优、自身最优的差值的乘积项,对算法中的速度-位置更新方程进行了改进,更好地保持粒子群的多样性,使算法可以不断地进化,提高算法的收敛速度,避免过早陷入局部最优解的现象。分别采用Sphere、Rosenbrock、Griewank和Rastrigrin4个标准函数对其进行测试。对比结果表明引入第3方随机项的PSO基本模型和收敛因子模型比传统模型的最优适宜度精度和优化收敛速度都有所提高。因此,改进的算法表现出更强的求解能力。
Based on Particle Swarm Optimization (PSO), the third-party random item is introduced: the product of random number, global optimum and self-optimal difference, and the velocity-location update equation in the algorithm is improved to better maintain the particle swarm The diversity of the algorithm can make the algorithm evolve constantly, improve the convergence speed of the algorithm and avoid the phenomenon of premature falling into the local optimal solution. Sphere, Rosenbrock, Griewank and Rastrigrin four standard functions were tested. The comparison results show that the PSO model and the convergence factor model of the third-party random item are better than the traditional model in both the optimal fitness and the convergence speed. Therefore, the improved algorithm shows a stronger ability to solve.