direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments



Nature-inspired algorithms in sanitary engineering: modelling sediment transport in sewer pipes
Zitatschlüssel Zounemat-Kermani_softcomputing_2021
Autor Zounemat-Kermani, Mohammad and Mahdavi-Meymand, Amin and Hinkelmann, Reinhard
Jahr 2021
ISSN 1433-7479
DOI 10.1007/s00500-021-05628-1
Journal Soft Computing
Monat Feb
Zusammenfassung In this study, two nature-inspired optimization techniques such as firefly algorithm (FA) and butterfly optimization algorithm (BOA) are combined with adaptive neuro-fuzzy inference system (ANFIS) and group method of data handling (GMDH) models for optimal prediction of the complex phenomenon of volumetric concentration of sediment (Cv) in sewer systems. Three different scenarios based on the methods of dimensional analysis and forward selection are implemented for determining the input structure of ANFIS, GMDH, and regression models (multiple linear regression, MLR; stepwise regression; SR) regarding 13 independent hydraulic and geometric input variables. Several statistic criteria including the root-mean-square error (RMSE), mean absolute error (MAE), Nash–Sutcliffe efficiency (NSE), index of agreement (AI), coefficient of determination (R2), and comprehensive synthesis index (SI) as well as Taylor diagram were used to further quantify simulating and predicting accuracies. In comparison with the regression models and two empirical equations, the results obtained by standard machine learning models (ANFIS and GMDH) were very promising. However, such integration of FA and BOA noticeably improved the performance of ANFIS (around 7% improvement in RMSE criterion) and slightly optimized the performance of GMDH (less than 1% improvement in RMSE criterion) in modelling the process of Cv prediction.
Link zur Publikation Download Bibtex Eintrag

Zusatzinformationen / Extras


Schnellnavigation zur Seite über Nummerneingabe