فهرس المقالات pollution in rivers.genetic algorithm

• حرية الوصول المقاله
  • صفحة الملخص
  • نص كامل
  
  1 - Prediction of Longitudinal Dispersion Coefficient of Pollution in Rivers Using a Modified of Neural Networksby Genetic Algorithm
  عباس پارسایی امیر حمزه حقی آبی امیر مرادی نژاد
  Reductionof Surface water quality and pollution in the environment is majorproblems. This issuewill become more important because the rivers are as a source for supplied for drinking water forpeople, industrial and agriculture. Prediction and modeling of hydraulic pheno أکثر

  Reductionof Surface water quality and pollution in the environment is majorproblems. This issuewill become more important because the rivers are as a source for supplied for drinking water forpeople, industrial and agriculture. Prediction and modeling of hydraulic phenomenon is one of themost importantactivities of Hydraulic Engineering. Neural network is one of the most usefulmethods of data processing which capable of modeling the complex relationships between inputand output. In this study, for prediction of the dispersion coefficient of pollution in rivers andthedevelopment of neural network (ANN) and empirical formulas wasstudied. Best accuracy ofthem is related to the Tavakollizadeh and Kashefipur, formula which its error index R 2 􀀠 0.77 .To increase in the perdition of the dispersion coefficient, the multi-layer perceptron (MLP) wasdeveloped. Training process and simulation MLP model was conducted in the Matlabsoftwareenvironment.To increase the performance of the MLP, genetic algorithm for training process isused. The results showed that the MLP are more accurate in comparison with otherempiricalequations.Using genetic algorithms for neural network training the neural networkmodel will further increase its accuracy about the 19 percent. تفاصيل المقالة