Affiliations: [a] High Magnetic Field Laboratory, Chinese Academy of Sciences, Hefei, Anhui, China | [b] University of Science and Technology of China, Hefei, Anhui, China | [c] College of Mechanical and Electrical Engineering, Anhui Jianzhu University, Hefei, Anhui, China
Correspondence:
[*]
Corresponding author: Song Chen, High Magnetic Field Laboratory, Chinese Academy of Sciences, Hefei, Anhui, China. E-mail: gengyun97@163.com.
Abstract: Building energy consumption prediction per month is an important content of building energy consumption management and company’s financial budget. BP neural network with parameter optimization, network optimized by mind evolutionary algorithm, network optimized by genetic algorithm, network optimized by particle swarm algorithm and network optimized by adaptive weight particle swarm algorithm are used to forecast the energy consumption. The optimal values of the learning rate and hidden layer node number are choosen. The characteristics of various kinds of optimization algorithm are compared. The neural network optimized by adaptive weight particle swarm algorithm is proved to be the most accurate in predicting energy consumption.
Keywords: BP neural network, optimization algorithm, energy consumption prediction
