摘要
养殖池塘中的溶解氧(DO)对水产品的生长和品质有着至关重要的作用。为了提高溶解氧预测的准确性和有效性,提出了一种基于集合经验模态分解(EEMD)和萤火虫算法(FA)优化支持向量机(SVM)的组合预测模型。首先,将DO时间序列通过集合经验模态分解为一组去除噪声的并相对稳定的子序列。接着,利用相空间重构(PSR)重建分解子序列,在相空间中用SVM对各子序列进行建模预测。然后,利用萤火虫算法对SVM的参数进行优化,建立基于SVM的预测模型,最后得到原始DO序列的预测值。为了获得未来24小时的预测结果,采用单点迭代法实现多步预测。仿真结果表明,所提出的EEMD-FA-SVM组合预测模型比FA-SVM、EEMD-FA-BP和EEMD-PSO-SVM等模型具有更好的预测效果,能够满足现代渔业养殖水质精细化管理的高需求。
The dissolved oxygen(DO)in the pond plays an important role in the growth and quality of aquatic products.In order to improve the prediction accuracy and effectiveness of DO,a combined prediction model based on ensemble empirical mode decomposition(EEMD)and firefly algorithm(FA)optimized support vector machine(SVM)is proposed.Firstly,the DO time series were decomposed into a group of relatively stable subsequence by ensemble empirical mode decomposition.Secondly,the decomposed subsequence was reconstructed by the phase space reconstruction(PSR),and the subsequence was modeled and predicted using the SVM in phase space.Finally,the parameters of SVM were optimized by using firefly algorithm to obtain the predicted value of original DO sequence.This paper used single point iterative method to achieve multi-step prediction,in order to obtain the forecasting result of 24 hours in the future.The simulation results demonstrate that the proposed combinatorial prediction model EEMD-FA-SVM has better prediction effect than FA-SVM,EEMD-FA-BP and EEMD-PSO-SVM,which is proven to be an effective way to predict DO,and it can meet the actual demand for fine management of water quality in modern fishery culture.
作者
刘晨
李莎
丛孙丽
朱正伟
LIU Chen;Li Sha;CONG Sun-Li;ZHU Zheng-wei(School of IOT Engineering,Wuxi Taihu University,Wuxi Jiangsu 214064,China;College of information science and Engineering,Changzhou University,Changzhou Jiangsu 213164,China)
出处
《计算机仿真》
北大核心
2021年第1期359-365,共7页
Computer Simulation
关键词
溶解氧预测
集合经验模态分解
萤火虫算法
支持向量机
单点迭代法
Prediction of dissolved oxygen
Ensemble empirical mode decomposition(EEMD)
Firefly algorithm(FA)
Support vector machine(SVM)
Single point iterative method
