Titre : A Sustainability-Oriented 4-Step Integrated Decision-Making Algorithm for Battery Selection: Integrating Expert Judgment and Bootstrapping Analysis
Résumé :
Sustainability plays an increasingly vital role in shaping policies, technologies, and individual behavior, particularly in the context of achieving the European Union’s Green Deal objectives. Among the sectors contributing significantly to greenhouse gas (GHG) emissions, transportation stands out as a primary concern. Electric vehicles (EVs) offer a promising solution for reducing emissions due to their lack of combustion engines and zero tailpipe emissions during operation. However, the environmental and societal impacts associated with their production, especially during the extraction of raw materials for batteries, raise important concerns. These include risks related to the environment, society, governance, and the long-term resource availability of critical raw materials.
In response to these challenges, this thesis proposes a comprehensive battery selection framework for EVs using a 4-step integrated Multi-Criteria Decision-Making (MCDM) approach. The methodology evaluates 12 battery alternatives based on a set of sustainability-related criteria and includes inputs from both decision-makers (DMs) and citizens, who represent the end users of EVs. Unlike previous studies that rely solely on DM or customer evaluations, this research incorporates public perspectives, recognizing that consumer preferences significantly influence sustainable adoption. The results highlight the sensitivity of the final battery selection to citizen preferences, emphasizing the importance of participatory decision-making in sustainable design.
To further investigate the robustness and accuracy of criteria weighting, the second part of the thesis introduces a bootstrapping applied to DEMATEL and AHP methods under a fuzzy environment. This phase analyzes how varying the number of DMs and the sensitivity of the linguistic evaluation scales affect the stability of the results across different numbers of criteria. Findings show that increasing the number of DMs reduces error in weight calculations and that less sensitive linguistic scales produce more consistent and convergent outcomes. Additionally, as the number of criteria increases, fewer DMs are required to reach reliable results, which demonstrates the efficiency and scalability of the proposed approach.
Jury :
- Rapporteurs :
- Prof. Evren Sahin LGI, CentraleSupelec
- Prof. Damien Trentesaux LAMIH, Université Polytechnique Hauts de France
- MdC (HDR) Gilles Philippot ICMCB, Université de Bordeaux
- Examinateurs :
- MdC Tijana Vulevic Université de Belgrade
- MdC Pascale Marangé CRAN, Université de Lorraine
- Prof. Éric Levrat CRAN, Université de Lorraine
- Invités:
- Research Scientist Sylvain Kubler SnT, University of Luxembourg
- MdC Tuncay Gürbüz Galatasaray University
- Directeurs de thèse :
- Prof. Éric Rondeau CRAN, Université de Lorraine
- MdC Alexandre Nominé IJL, Université de Lorraine