混沌神经网络具有全局搜索能力 ,但其运用至今主要局限于组合优化 .通过对普通 Hopfield优化网络引入混沌噪声退火过程 ,提出了一种用于约束非线性全局优化的混沌退火神经网络 ,它易于实现 ,原理简明 ,应用广泛 .对很复杂的测试函数的数字试验表明 ,该模型能够高效、可靠地搜索到全局最优 ,其性能超过遗传算法 GAMA S Chaotic neural network has the ability of global search, but its application has been mainly limited to combinatorial optimization so far.By introducing chaotic noise annealing into ordinary Hopfield networks, a chaotic annealing neural network for constrained nonlinear global optimization is proposed, which is easy to implement , The principle is concise and the application is extensive.Numerical tests on very complicated test functions show that this model can search the global optimum efficiently and reliably and its performance exceeds the GAMA S