An integer linear programming approach for pavement maintenance and rehabilitation optimization

A highway in poor conditions can raise transportation costs. Due to budgetary constraints, pavement maintenance programming is considered a difficult decision-making problem. In this article we propose a novel mathematical model and a different variant of the pavement maintenance management problem,...

Full description

Saved in:
Bibliographic Details
Published in:The international journal of pavement engineering 2022-07, Vol.23 (8), p.2710-2727
Main Authors: Gomes Correia, Matheus, Bonates, Tibérius de Oliveira e, Prata, Bruno de Athayde, Nobre Júnior, Ernesto Ferreira
Format: Article
Language:English
Subjects:
Benchmarks
Decision making
Greedy algorithms
integer linear programming
Integer programming
Linear programming
Maintenance management
Mathematical models
network-level maintenance
Optimization
optimization model
pavement maintenance and rehabilitation
Pavement management
Pavements
Performance indices
Rehabilitation
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A highway in poor conditions can raise transportation costs. Due to budgetary constraints, pavement maintenance programming is considered a difficult decision-making problem. In this article we propose a novel mathematical model and a different variant of the pavement maintenance management problem, solved with integer linear programming. The novelty of this approach is the use of the Pavement Surface Rating as the condition indicator, along with a proposed conversion strategy between most used performance indices. Additionally, we propose a simpler and broader deterioration model, when compared to existent ones, using a table system. This renders the model to be solved easily, allowing it to be implemented worldwide, given its generic characteristics. Many computational experiments were performed, both on artificial benchmark instances and on a real-world case study. The proposed model is shown to obtain optimal solutions in short computational times, and it is able to solve much larger instances than the ones found in the literature. Optimal solutions from benchmark instances, consisting of 5,000 segments and an analysis period of 30 years, were found in less than 45 minutes. Additionally, the optimal solutions have a difference of more than 20% in average, when compared to a greedy algorithm.
ISSN:1029-8436
1477-268X
DOI:10.1080/10298436.2020.1869736