Abstract: In order to prevent appearing premature convergence in searching iterations of particle swarm optimization algorithm, an adaptive chaotic embedded particle swarm optimization algorithm is proposed. It embeds the chaos variables into the standard PSO algorithm（SPSO） and adjusts parameters adaptively. In this algorithm, to take full advantage of the randomicity, ergodicity and disciplinarian of chaos, Tent chaotic maps is used to substitute the random numbers of the SPSO; the inertia weight and acceleration coefficient is adjusted adaptively with nonlinearly strategy; the population fitness variance of particle swarm and average distance amongst points are put forward to estimate particles whether being focusing or discrete, then chaotic researching is applied to jump out of local optimum. Simulation experimental results show this algorithm has very good global optimization ability, high precision and robustness.

Key words: embedded particle swarm optimization algorithm, chaos, adaptive, tent mapping, average distance amongst points, fitness variance

摘要： 为避免粒子群算法后期出现早熟收敛，提出一种基于Tent映射的自适应混沌嵌入式粒子群算法。将混沌变量嵌入到标准粒子群算法中，且对参数进行自适应调整。算法采用Tent映射生成的混沌序列来取代基本粒子群算法中的随机数，充分利用了混沌运动的随机性、遍历性和规律性；惯性权重和学习因子采用非线性的自适应调整策略；建立平均粒距与适应度方差相结合的早熟收敛判断机制，并且以混沌搜索的方式来跳出局部最优。测试函数仿真结果表明，该算法具有良好的全局搜索能力，寻优精度较高，鲁棒性好。

关键词: 嵌入式粒子群算法, 混沌, 自适应, 帐篷映射, 平均粒径, 适应度方差