DocumentCode :
2146431
Title :
The Application of Ant Colony Algorithm in Combined Power Load Forecasting
Author :
Zhiping, Fan ; Tiansheng, Hong
Author_Institution :
Coll. of Comput. Sci. & Educ. Software, Guangzhou Univ., Guangzhou
fYear :
2008
fDate :
30-31 Dec. 2008
Firstpage :
90
Lastpage :
93
Abstract :
The ant colony algorithm also named ACO method, Artificial ants have memory function, it in the process of movement through a pheromone path to release, finally found a road from their nests to food source through the shortest, called the catalytic processes of ants, which is a positive feedback mechanism, it has good robustness, fast convergence speed, easy to get the global optimal solution. Based on random increase and nonlinear wave -- residual series, grey prediction can reflect the increasing and support vector machine can show the nonlinear relationship. The improving ACO method can make the optimization weight to achieve the goal of accuracy, consistency to the prediction values, finally precision of the series can be improved obviously. Through computation of power load in a province, the experimental results show that this method can greatly improve the accuracy of the load forecasting.
Keywords :
load forecasting; optimisation; power engineering computing; support vector machines; ant colony algorithm; artificial ants; catalytic processes; combined power load forecasting; global optimal solution; grey prediction; memory function; nonlinear wave; positive feedback mechanism; residual series; support vector machine; Application software; Computer science; Information technology; Iterative algorithms; Load forecasting; Predictive models; Risk management; Roads; Software algorithms; Support vector machines; 1) method; Ant colony algorithm; GM (1; combined power load forecasting;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
MultiMedia and Information Technology, 2008. MMIT '08. International Conference on
Conference_Location :
Three Gorges
Print_ISBN :
978-0-7695-3556-2
Type :
conf
DOI :
10.1109/MMIT.2008.127
Filename :
5089066
Link To Document :
بازگشت