论文部分内容阅读
针对水电站水库常用的调度规则所存在的问题，进行了维护性讨论，探索了一种对知识规则的更新方法，以便于适应水库调度的实际要求。更新方法是以并行神经网络结构为基础，由知识增长的追加学习算法组成。经新增样本３×３６个的仿真应用，检验了该方法的可行性。
In view of the problems existing in the commonly used scheduling rules of hydropower stations, a maintenance discussion was conducted and an updated method of knowledge rules was explored to adapt to the actual requirements of reservoir dispatching. The update method is based on parallel neural network structure and consists of additional learning algorithm of knowledge growth. The new sample 3 × 36 simulation applications, test the feasibility of the method.