Risk-adjusted economic life modeling of large-scale municipal assets using hierarchical machine learning

Yersultan Tulebayev, Айвар Сахипов, Dametken Baigozhanova, Shynar Yelezhanova, Gulash Kochshanova

2026

Full text Open access: Gold Metadata (CSL-JSON)
DOI 10.47533/2026.1606-146x.1-13
Data confidence 85 high
How this score is calculated

It is built from checkable properties of the record itself: whether the article has a DOI, how many independent sources confirmed its data, whether the required fields are filled in, and how recent the latest confirmation is. Disagreement between sources and a missing identifier lower it. A value of 80 or above means the record was published without further checking; 60 to 79 means it is published at reduced confidence, with a manual check advisable. A record verified by an editor by hand is never pushed below 90 by a recalculation. This scores our data about the article, not the article itself.

Abstract

Municipal and public-sector organizations manage extensive portfolios of physical assets whose aging, increasing maintenance costs, and heterogeneous failure risks pose significant challenges for sustainable infrastructure management. Classical Economic Life (EL) models determine optimal replacement timing by minimizing equivalent annual costs under deterministic assumptions, but they do not explicitly account for asset-level uncertainty and probabilistic failure risk (Hritonenko & Yatsenko, 2005; Hritonenko & Yatsenko, 2006; Yatsenko & Hritonenko, 2011). In parallel, recent advances in machine learning enable large-scale risk prediction for infrastructure assets; however, such models are often weakly connected to formal economic decision frameworks (Ji et al., 2025; Alshboul et al., 2022; Mistry & Hough, 2024; Snider & McBean, 2021; Jensen et al., 2025, Tulebayev, Mukhanova & Goranin, 2025, Munoz-Villamizar, Rafavy & Casey, 2022; Mantas, Abellán & Castellano, 2019; Ismail, Fathi & Zayed, 2018). This paper proposes a Risk-Adjusted Economic Life (RA-EL) framework that integrates a modified Economic Life model with a hierarchical hybrid machine learning architecture for municipal asset renewal planning. Using a real-world dataset comprising more than 600,000 municipal assets, failure risk is estimated via a stacking ensemble and incorporated into an Equivalent Expected Annual Cost (EEAC) formulation. The proposed approach embeds probabilistic risk into replacement theory, improving replacement prioritization and supporting risk-informed capital investment decisions in large-scale municipal asset portfolios.