TY - JOUR N2 - Accurate prediction of power load plays a crucial role in the power industry and provides economic operation decisions for the power operation department. Due to the unpredictability and periodicity of power load, an improved method to deal with complex nonlinear relation was adopted, and a short-term load forecasting model combining FEW (fuzzy exponential weighting) and IHS (improved harmonic search) algorithms was proposed. Firstly, the domain space was defined, the harmony memory base was initialized, and the fuzzy logic relation was identified. Then the optimal interval length was calculated using the training sample data, and local and global optimum were updated by optimization criteria and judging criteria. Finally, the optimized parameters obtained by an IHS algorithm were applied to the FEW model and the load data of the Huludao region (2013) in Northeast China in May. The accuracy of the proposed model was verified using an evaluation criterion as the fitness function. The results of error analysis show that the model can effectively predict short-term power load data and has high stability and accuracy, which provides a reference for application of short-term prediction in other industrial fields. L1 - http://journals.pan.pl/Content/114131/PDF/14_AEE-2019-4_INTERNET.pdf L2 - http://journals.pan.pl/Content/114131 PY - 2019 IS - No 4 EP - 923 DO - 10.24425/aee.2019.130691 KW - evaluation criteria KW - exponential fuzzy time series KW - fitness function KW - improved harmony search algorithms KW - load forecasting KW - optimal interval length A1 - Yu, Mingxing A1 - Zhu, Jiazheng A1 - Yang, Li PB - Polish Academy of Sciences VL - vol. 68 DA - 2019.12.02 T1 - Short-term load prediction model combining FEW and IHS algorithm SP - 907 UR - http://journals.pan.pl/dlibra/publication/edition/114131 T2 - Archives of Electrical Engineering ER -