EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

3426. Performance comparison of machine learning models in detecting the best-performing methods for the resource-constrained single-and multi-project scheduling problem

Invited abstract in session TC-6: Advancements of OR-analytics in statistics, machine learning and data science 14, stream Advancements of OR-analytics in statistics, machine learning and data science.

Tuesday, 12:30-14:00
Room: 1013 (building: 202)

Authors (first author is the speaker)

1. Weikang Guo
Computer science and technology, KTH, Ghent University
2. Mario Vanhoucke
Faculty of Economics and Business Administration, Ghent University, Vlerick Business School, University College London
3. José Coelho
Department of Sciences and Technology (DCeT), Universidade Aberta

Abstract

Priority rules are the most used methods in many commercial software tools for scheduling projects under limited resources because of their ease of implementation, intuitive working, and fast speed. The branch-and-bound (B&B) procedure is the most common way to deal with the resource-constrained project scheduling problem to find the optimal solution in a finite number of steps. Meta-heuristics are the most effective methods to solve project scheduling. Although these methods have been developed, no single method has been shown to be the best. To fill this gap, we rely on machine learning classification and ranking models to find the relation between the characteristics of the project instances and the performance of solution methods.

This study not only aims to compare the performance of different machine learning models but also proves that machine learning can be used to solve the resource-constrained multi-project scheduling problem (RCMPSP). The contribution is threefold: First, we extend two types of machine learning models to solve the multi-project scheduling problem. Second, we extend both machine learning models to priority rules, branch-and-bound procedures, and meta-heuristics, respectively. Third, we compare the performance of two types of machine learning models on the three categories of solution methods. An extensive computational experiment is set up to compare the performance of machine learning models with the existing methods from the literature.

Keywords

Status: accepted

Back to the list of papers