提出了一种RBF网非线性动态系统在线建模的资源优化网络(RON)方法．RON

在资源分配网络的学习过程中引入了滑动窗口和网络结构在线优化的思想，使网络能根据最

近一段时间内的误差信息自动实现网络结构优化，从而使RBF网既能在线适应对象的变化，

又能使网络规模维持在较小水平，并保证了网络的泛化能力．使用滑动窗口技术使RON对学

习参数变化具有较好的鲁棒性，并更易收敛．三个标准例子演示了算法的有效性．-presents a RBF network nonlinear dynamic system modeling online resource optimization network (RON) method. RON resources distribution network in the learning process of introducing a sliding window and on-line optimization of the network structure of thinking, so that the network can According to the most recent information within the error automatically optimizing the network structure, so that the RBF network can meet the targets of the online changes also enable the network to maintain the small-scale level, and to ensure that the network generalization ability. Use sliding window technology allows RON learning parameters is robust, and easier to convergence. 3 standard examples demonstrate the effectiveness of the algorithm.