摘要
通过在神经网络状态空间演化方程中引入一个非线性反馈项，使神经网络系统的动力学表现出混沌特性．为将混沌动力学作为搜索机制应用于优化问题，又引入一个调节机制构成了暂态混沌神经网络模型．本文着重分析了暂态混沌神经网络动力学行为，并将其应用于旅行推销员问题．实现了全局优化且有较快的收敛速度．
The dynamics of neural network system will present chaotic characters through adding a nonlinear feedback term in the state evolution equation. In order to use the chaotic dynamics as the heuristic mechanism in the optimization problems, we also introduced an adjusting method to construct a transiently chaotic neural network model. In this paper, we mainly analyzed the dynamical behaviour of the transiently chaotic neural network system, and applied the model to the traveling salesman problem. In the application, global optimization is realized with faster convergence speed.
出处
《南开大学学报（自然科学版）》
CAS
CSCD
北大核心
1999年第3期99-103,共5页
Acta Scientiarum Naturalium Universitatis Nankaiensis
基金
国家九五攀登计划非线性科学项目资助
关键词
神经网络
暂态混沌动力学
组合优化
TSP
neural network
transient chaotic dynamics
combinatorial optimization
TSP
