一种新型短期负荷预测模型的研究及应用

Research and application of a new forecasting method for short-term load

提出基于量子神经网络进行短期负荷预测的方法。建立起由量子比特神经元改进而来的三层量子神经网络模型,应用量子门进行计算,输出层的权值和阈值进行非线性变换,实现神经网络非线性映射功能,并利用遗传算法优化了量子神经网络的初始参数。仿真结果表明,此算法在收敛速度,泛化能力和预测精度方面明显优于普通三层BP网络。

This paper proposes the forecasting method for short-term load based on QNN,establishes the three-layer neural network improved from the quantum bit neuron,realizes non-linear mapping of neural network through the transformed weights and threshold value of output layer worked out by quantum gate,and optimizes the initial parameters of quantum neural network by genetic algorithm.Simulation result proves that this algorithm is superior to the common three-layer BP network from the aspects of the rate of convergence,generalization ability and forecasting accuracy.