基于结合混沌纵横交叉的粒子群算法优化极限学习机的短期负荷预测

为了解决传统的单一负荷预测模型精度低以及常规智能算法在解决高维、多模复杂问题时容易陷入局部最优的问题，提出了一种结合混沌纵横交叉的粒子群算法（CC-PSO）优化极限学习机(ELM)的短期负荷预测模型。ELM的泛化能力与其输入权值和隐含层偏置密切相关，采用结合混沌纵横交叉的粒子群算法优化ELM的输入权值与隐含层偏置，提高了ELM的泛化能力和预测精度。选择广东某地区实际电网负荷数据进行分析，研究结果表明，相对于BP神经网络和支持向量机，ELM具有更高的泛化能力和预测精度；CC-PSO相对于粒子群和遗传算法具有更高的全局搜索能力，CC-PSO-ELM模型具有较高的负荷预测精度。

Short-term Load Forecasting based on Extreme Learning Machine Optimized by Particle Swarm Optimization integrated with Chaotic Crisscross Optimization

To solve the problem that the forecasting accuracy of traditional single load forecasting model is low and the conventional intelligent algorithm is likely to be trapped into the local optimal problem when solving the high-dimensional and multimodal optimization problems, this paper presented a model which used particle swarm optimization integrated with chaotic crisscross optimization(CC-PSO) to optimize extreme learning machine(ELM) for short-term load forecasting. The generalization ability of ELM is closely related to its input weights and hidden layer biaes, using CC-PSO to optimize the input weights and hidden layer biaes of ELM could improve the generalization ability of ELM and forecasting accuracy. The actual power grid load data of a certain area in Guangdong was selected and analyzed, the results show that ELM has higher generalization ability and prediction accuracy compared with BP neural networks and support vector machines; CC-PSO has higher global search capability relative to particle swarm optimization and genetic algorithm; and CC-PSO-ELM model has higher forecasting accuracy.