EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

4358. A hybrid genetic algorithm for non-regular flexible job shop arising from deconstruction applications

Invited abstract in session WD-28: Advancements of OR-analytics in statistics, machine learning and data science 11, stream Advancements of OR-analytics in statistics, machine learning and data science.

Wednesday, 14:30-16:00
Room: 065 (building: 208)

Authors (first author is the speaker)

1. Corentin Juvigny
Information Systems, Decision sciences and Statistics, ESSEC Business School

Abstract

The field of building deconstruction is a significant contributor to waste generation. However, despite the availability of many waste recovery techniques to valorize this waste, much of it is currently lost due to poor waste stream management.
This study aims to propose a modeling of the deconstruction sector that optimizes dismantling operations and their associated waste flows to raise the recovery rate of the waste, thereby lowering both the financial and environmental costs of the field. This model of the deconstruction sector is based on a bi-level problem. Its upper-level part encompasses the scheduling of operations. The latter hinges on a weighted flexible job shop problem (wFJSP) with release dates and is assessed by a non-regular criterion.
We propose a new hybrid disjunctive graph genetic algorithm (HDGGA) for wFJSP, integrating into a genetic algorithm, based on a new encoding, a local search approach employing a novel extension of the disjunctive graph representation that takes account of the idle machine times.
The lower-level problem models the allocation of the waste induced.
Experiments are carried out on two sets of instances, including one derived from data collected amongst the actors of the field working in the Lille metropolis.
The first results exhibit the interest in the genetic-based hybrid method proposed, indicating that it outperforms the existing genetic algorithm-based approach, especially in medium-sized and large-sized instances.

Keywords

Status: accepted

Back to the list of papers