Short-term Electric Load Forecasting Method Based on Improved Genetic Algorithm Optimizing Extreme Learning Machine

Abstract

The invention discloses a short-term power load prediction method based on an improved genetic algorithm to optimize an extreme learning machine. The hill-climbing method is used to select the best in the offspring population again, select an initial individual, select another individual in the adjacent area, and combine the two The fitness values ​​of each individual are compared, and individuals with good fitness values ​​are left; if the initial individual is replaced or a better individual cannot be found in several iterations, the iteration is stopped, and the search direction of the genetic algorithm is adjusted by the hill-climbing method. Make the genetic algorithm move towards the global optimal direction, obtain the optimal weights and thresholds, obtain the network optimization prediction model, and compare and analyze the prediction results of the BP network and extreme learning machine, including the selection of the input and output of the prediction network model , Improved genetic algorithm to optimize the algorithm of extreme learning machine and analysis of prediction results. The invention has faster training speed and more accurate prediction results, and is suitable for modern short-term power load prediction with numerous influencing factors and huge data volume.

Description

Technical field: [0001] The invention relates to a short-term power load forecasting method based on an improved genetic algorithm to optimize an extreme learning machine, which is an important content in the economic dispatch of a power system and an important module of an energy management system (EMS). Background technique: [0002] The traditional extreme learning machine (Extreme Learning Machine, ELM) randomly generates input layer weights and hidden layer thresholds, which makes the network unstable and prone to overfitting, which affects the generalization performance of the network. Traditional Genetic Algorithm (GA) has limited space search capability, and it is easy to converge to a local optimal solution, resulting in premature problems. Invention content: [0003] The technical problem to be solved by the present invention is to provide a short-term power load forecasting method based on the improved genetic algorithm to optimize the extreme learning machine f...

AI Technical Summary

Problems solved by technology

[0003] The technical problem to be solved by the present invention is to provide a short-term power load forecasting method based on the improved genetic algorithm to optimize the extreme learning machine for the problem that the input layer weights and hidden layer thresholds randomly generated by ELM lead to network instability. The improved genetic algorithm, that is, the improved genetic algorithm (IGA) optimizes the weights and thresholds of the ELM, thereby improving network performance and improving the accuracy of network short-term load prediction; applying the improved genetic algorithm to optimize the extreme learning machine (IGA-ELM) for short-term load Forecasting can speed up the learning speed of the forecasting network, enhance the stability of the forecasting network, and improve the prediction accuracy of the forecasting network

Images