EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

3211. Optimizing the usage of cut and fill materials in a large-scale railway construction project

Invited abstract in session WA-52: Models and algorithms for real-life combinatorial optimization problems, stream Combinatorial Optimization.

Wednesday, 8:30-10:00
Room: 8003 (building: 202)

Authors (first author is the speaker)

1. Felipe Lemos
Universidade Federal da Paraíba
2. Bruno Bruck
Federal University of Paraiba
3. Maria Battarra
4. Anand Subramanian
Universidade Federal da Paraíba

Abstract

In this work, we investigate the problem of reusing dug materials in a large-scale railway building construction project. The project is inspired by a real-world scenario from HS2, the company responsible for building a 225-kilometer fast train line in the UK. The company estimates that the dug material will amount to around 132 million tons, and material imported for construction will be around 20 million tons. The dug material can often be repurposed for fill operations to minimize environmental impact. In this research, we model the problem of allocating dug materials to fill operations, considering operational and financial constraints. We propose two approaches to solve the problem. The first one consists of an integrated MILP model, while the second one is a heuristic algorithm that initially solves a simplified version of the integrated model for each month of the time horizon, and then calls another MILP that tries to find a feasible solution. Both approaches were tested on instances generated based on real data provided by the company. Preliminary results show that the first approach is only capable of finding optimal solutions for small and medium sized instances. On the other hand, the second one is significantly more efficient in terms of runtime required to find feasible solutions, especially for larger instances, but it struggled to find the optimal solutions obtained by the first approach.

Keywords

Status: accepted

Back to the list of papers