共生进化免疫神经网络在电力系统短期负荷预测中的应用

Application of symbiotic evolutionary immune neural network in short-term load forecasting of power system

为了克服传统BP神经网络在结构设计和学习算法中存在的缺陷 ,提出了一种共生进化免疫神经网络来预测电力系统短期负荷。其中利用共生进化原理设计神经网络 ,通过对神经元群体进行优化设计 ,显著地减轻了计算量。在进化过程中 ,结合免疫算法中的浓度机制和个体多样性保持策略进行免疫调节 ,有效地克服了未成熟收敛现象 ,提高了群体的多样性 ,加快了网络设计速度。算例计算表明 。

To overcome the defects in structure design and learning algorithm of traditional BP neural network, a symbiotic evolutionary immune neural network for short-term load forecasting of power system was put forward, which used the symbiotic evolutionary principle to design the neural network and thus dramatically reduces the amount of computation through optimal designing of the neutron population. During the evolving process of the neutron population, combining the concentration mechanism of immune algorithm and the strategy of keeping the individual diversity to perform immune adjustment, the phenomenon of premature convergence is overcome effectively, the diversity of the cluster is improved and the speed of network designing is accelerated. Calculation example shows that the method has the advantage of shorter training period and higher forecast accuracy.