
脉冲暂态混沌神经网络在约束非线性规划中的应用
李旲1,2、曹宏铎3、胡云昌2、山秀明1
Pulse Transiently Chaotic Neural Network for Nonlinear Constrained Optimization
Li Ying1,2、 Cao Hongzhao3、 Hu Yunchang2、 Shang Xiuming1
脉冲暂态混沌神经网络(PTCNN)是对暂态混沌神经网络的改进,呈现丰富的动力学性质,具有很强的跳出局部最小点的功能,在解决无约束非线性规划问题时,可以找到包括全局和局部最小值的尽量全面的最优解。当遇到带约束条件的非线性规划问题时,只有对约束条件进行合理处理,才能更有效地解决约束非线性规划问题。文章使用惩罚函数方法对含有约束条件的非线性规划问题进行处理,将其变成一个不含约束条件的非线性规划问题,进而用PTCNN求解,得到了令人满意的结果。
Pulse Transiently Chaotic Neural Network (PTCNN) can find almost all optima including the part optima and the global with its abundance dynamical characteristic, when is used in nonlinear non-constrained optimization. The optimization problem is first unconstrained by virtue of non-differentiable exact penalty function, and is further solved by PTCNN. It is showed by an example that this method is efficient.
PTCNN / penalty function / nonlinear constrained optimization
/
〈 |
|
〉 |