AN ALGORITHM OF CHAOTIC DYNAMIC ADAPTIVE LOCAL SEARCH METHOD FOR ELMAN NEURAL NETWORK
AN ALGORITHM OF CHAOTIC DYNAMIC ADAPTIVE LOCAL SEARCH METHOD FOR ELMAN NEURAL NETWORK
复制标题
DOI：
--
复制
发表时间：
2010
期刊：
影响因子：
--
通讯作者：
Zhiqiang Zhang;Zheng Tang;Shangce Gao;Gang Yang
中科院分区：
文献类型：
--
作者：
Zhiqiang Zhang;Zheng Tang;Shangce Gao;Gang Yang
在本文中，我们提出了一种用于预测与太阳黑子相关的时间序列的高效算法，即混沌动态自适应局部搜索（CDALS）算法。该算法基于对部分递归的埃尔曼神经网络（ENN）的利用，并且可分为两个主要步骤：第一个是我们先前工作中提出的自适应局部搜索（ALS）的基本模型。之后，通过将混沌信号引入ALS，提出了一种混合局部搜索方法。因此，ALS和混沌被混合以形成一种强大的CDALS算法，它合理地结合了ALS的搜索能力和混沌搜索行为。仿真结果表明，CDALS算法能够以较高的概率最终达到全局最优或其良好近似值，在合理的迭代次数内有效地提高了搜索效率和质量。
In this paper, we present an efficient algorithm for the prediction of sunspotrelated time series, namely the Chaotic Dynamic Adaptive Local Search (CDALS) algorithm. This algorithm is based on exploiting partially recurrent Elman Neural Network (ENN) and it can be divided into two main steps: the first one is the basic model of the Adaptive Local Search (ALS) proposed in our previous work. After that, a hybrid local search method is proposed by introducing the chaos signals into ALS. Thus, ALS and chaos are hybridized to form a powerful CDALS algorithm, which reasonably combines the searching ability of ALS and chaotic searching behavior. Simulation results show that the CDALS algorithm can eventually reach the global optimum or its good approximation with high probability, effectively enhance the searching efficiency and quality within reasonable number of iterations.