Estimation of a logistic regression model by a genetic algorithm to predict pipe failures in sewer networks

Alicia Robles-Velasco (), Pablo Cortés, Jesús Muñuzuri and Luis Onieva
Additional contact information
Alicia Robles-Velasco: ETSI. Universidad de Sevilla
Pablo Cortés: ETSI. Universidad de Sevilla
Jesús Muñuzuri: ETSI. Universidad de Sevilla

OR Spectrum: Quantitative Approaches in Management, 2021, vol. 43, issue 3, No 7, 759-776

Abstract: Abstract Sewer networks are mainly composed of pipelines which are in charge of transporting sewage and rainwater to wastewater treatment plants. A failure in a sewer pipe has many negative consequences, such as accidents, flooding, pollution or extra costs. Machine learning arises as a very powerful tool to predict these incidents when the amount of available data is large enough. In this study, a real-coded genetic algorithm is implemented to estimate the optimal weights of a logistic regression model whose objective is to forecast pipe failures in wastewater networks. The goal is to create an autonomous and independent predictive system able to support the decisions about pipe replacement plans of companies. From the data processing to the validation of the model, all stages for the implementation of the machine-learning system are explored and carefully explained. Moreover, the methodology is applied to a real sewer network of a Spanish city to check its performance. Results demonstrate that by annually replacing 4% of pipe segments, those whose estimated failure probability is higher than 0.75, almost 30% of unexpected pipe failures are prevented. Furthermore, the analysis of the estimated weights of the logistic regression model reveals some weaknesses of the network as well as the influence of the features in the pipe failures. For instance, the predisposition of vitrified clay pipes to fail and of that pipes with smaller diameters.

Keywords: Logistic regression; Binary classifier; Pipe failures; Genetic algorithm; Sewer networks (search for similar items in EconPapers)
Date: 2021
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
http://link.springer.com/10.1007/s00291-020-00614-9 Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:orspec:v:43:y:2021:i:3:d:10.1007_s00291-020-00614-9

Ordering information: This journal article can be ordered from
http://www.springer. ... research/journal/291

DOI: 10.1007/s00291-020-00614-9

Access Statistics for this article

OR Spectrum: Quantitative Approaches in Management is currently edited by Rainer Kolisch

More articles in OR Spectrum: Quantitative Approaches in Management from Springer, Gesellschaft für Operations Research e.V.
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().