Dissolved Oxygen Prediction in Apostichopus Japonicus Aquaculture Ponds by BP Neural Network and AR Model
作者单位：College of Information and Electrical EngineeringChina Agricultural University
会议日期：2010年
学科分类：090801[农学-水产养殖] 12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 0908[农学-水产] 081104[工学-模式识别与智能系统] 08[工学] 09[农学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）]
基 金：financially supported by the National High Technology Research and Development Program of China(2007AA10Z238) Beijing Natural Science Foundation(4092024) The Ministry of Science and Technology of the People’s Republic of China for their financial support(2006BAD10A02-05)
关 键 词：DO Concentration Prediction BP Neural Network AR Model Sea Cucumber
摘 要：In the process of rearing sea cucumber Apostichopus japonicus,Dissolved Oxygen（DO） concentration adjustment and control is the key point to sea cucumber *** paper proposed both traditional time series method（auto regressive（AR） model） and BP neural network for the DO concentration forecasting in sea cucumber aquaculture *** aim to make four different interval length（0.5 hour,1 hour and 2 hours） prediction of DO concentration in sea cucumber breeding *** DO concentration is the result of many factors,such as water temperature,solar radiation,water level,air temperature,PH and EC value of water,*** different factors have complicated nonlinear relations which are difficult to describe with simple mathematical formula.A BP neural network using water temperature,past time DO concentration and EC value as inputs to make 2 hours ahead DO concentration prediction was constructed in this *** model was also employed to forecast the DO *** show that predictive values of both models are accurate according to the observed *** AR model has better performance in short interval prediction and BP neural network does better in longer interval *** models demonstrated potential for use as an intensive aquaculture management and forecasting tool to predict the near-term DO concentrations.