暂态混沌神经网络及其在优化问题中的应用研究

Transient Chaotic Neural Netwgork and Its Optimization of the Applied Research

为了分析研究暂态混沌神经网络特性及其优化机制,在分析与研究暂态混沌神经元模型基础上,通过在Matlab软件中编程仿真分析,比较神经网络的动力学特性及各参数对于网络的寻优过程影响。暂态混沌神经网络模型利用混沌所固有的随机性和轨道遍历性,在大范围内按其自身规律进行搜索,搜索过程按混沌轨道遍历,不受目标函数限制,从而具有克服陷入局部极小的能力可有效地解决一系列组合优化问题。这里根据网络动力学特性合理选择控制网络参数,通过仿真很好地解决了非线性函数优化问题和10个城市的TSP问题。相对于传统参数选择依靠经验使优化结果更具说服力,优化结果令人满意。从而有利于这种混沌神经网络在优化问题中的推广。

For analysing and studing the characteristic and optimized mechanism of transient chaotic neural network,based on analysis of transient chaotic neuron model, through programming the simulation analysis in the Matlab software to compare the neural network dynamic characteristic and various parameters regarding the network optimization process influence. Transient chaotic neural network model by the chaos inherent in the use of random traversal of the track,and carries on the search in wide range according to its own rule, the search process, according to traverse chaotic orbit, free from restrictions on the objective function,which has overcome the local minimum The ability to effectively can solve a series of combinatorial optimization problems. The control network parameter is selected according to the network dynamic characteristic, through the simulation,problems of non - linear function optimization and 10 city TSP problems are solved. The optimization results is satisfied. And thus is conducive to such a chaotic neural network optimization problem in the promotion.