变参数细胞神经网络的分数阶可切换多元电路设计及仿真

构建新的整数阶三维细胞神经网络系统，通过对其非线性动力学分析、数值计算与电路仿真，验证了该系统混沌吸引子的存在性及物理上的可实现性。同时通过调节线性参数b，研究了新的细胞神经网络系统在基于分数阶qi（i＝1，2，3）不同组合条件下所表现出的混沌特性。结合分数阶电路理论分析分数阶电路中各单元电路形式，并设计了相应参数b可变、阶数值qi可切换的分数阶细胞神经网络电路系统。经统计本设计可实现13824种多元组合电路，并选取具有代表性组合电路进行电路仿真。仿真结果表明，多元电路仿真和数值仿真具有相似的混沌相图，从而证实了细胞神经网络在分数阶条件下仍表现出丰富的动力学特性，具有灵活实用价值和现实推广意义。

Mu lti-Ele ment Circu it Si mu lation of Alterable Para meters and Switchable Fractional-Order Cellular Neural Networks

Designed a new integer-order and three-dimensional cellular neural networks system.Through nonlinear dy-namics analysis,numerical calculation and circuit simulation,the existence of chaotic attractor and realizability in physical a-bout the system were verified.At the same time,studied the system’s chaotic characteristics under the conditions of different combination about fractional order qi(i=1,2,3),by adjusting the linear parameter b.According to the theoretical analysis of each unit circuit of fractional-order circuit,designed a corresponding fractional order cellular neural networks system,which has variable parameter b and switchable qi.Statically,the design can realize 13824 kinds of multiple combination circuits.This pa-per selects several representative combinational circuits for circuit simulation.The simulation results show that multiple circuit simulation and numerical simulation have similar chaos phase diagram,which confirmed the system still has rich dynamic char-acteristics under the condition of the fractional-order,also has flexible application value and realistic significance.