Invited Speakers

Fouad Ben Abdelaziz, Distinguished Professor at NEOMA BS, Rouen Campus, France.
Chair of the first AFROS conference in Tunisia.
Fouad.BEN.ABDELAZIZ@neoma-bs.fr
https://neoma-bs.com/professors/ben-abdelaziz-fouad-2/
The Art of Mathematical Modeling: This presentation, discusses the art and science of mathematical modeling, tracing the path from decision-making processes to optimization. It highlights the role of mathematical models in selecting the best alternatives from a set of options, considering the challenges posed by non-continuous functions and the need for conditions that ensure the existence and uniqueness of optimal solutions. The presentation also explores discrete multi objective optimization problems, and addresses the computational complexities involved. Through the lens of rationality, uncertainty, it examines how mathematical models can handle imperfect information and dynamic situations. The discussion extends to the interaction between multiple decision-makers, cooperation versus. Ultimately, the presentation underscores the complexity and nuances of real-life problems, illustrating how mathematical models, while powerful, must be carefully crafted and interpreted within their limitations.

A. Ridha Majhoub
Department of Statistics and Operations Research, Kuwait University, Kuwait
LAMSADE, Université Paris-Dauphine, Paris, France
ridha.mahjoub@ku.edu.kw , ridha.mahjoub@lamsade.dauphine.fr
https://www.lamsade.dauphine.fr/~mahjoub/

Network Design and Cutting Plane Algorithms: For the past few decades, combinatorial optimization techniques have shown to be powerful tools for formulating, analysing and solving problems arising from practical decision situations. In particular, cutting plane techniques have been successfully applied to many well- known NP-hard problems. The equivalence between separation and optimization over a polyhedron, and the evolution of computational tools have been behind the big development of these methods. The so-called Branch-and-Cut method, which is inspired from this equivalence, is now widely used for obtaining optimal and near-optimal solutions for hard and large sized combinatorial optimization problems. In this talk, we present these methods and discuss some applications to network design problems.

El-Ghazali Talbi
CRISTAL – University of Lille & INRIA, France
el-ghazali.talbi@univ-lille.fr

How optimization can help to design deep neural networks : In recent years, research in metaheuristic optimization approaches in the automatic design and configuration of deep neural networks has become increasingly popular. Although various approaches have been proposed, there is a lack of a unified optimization framework for this hot research topic. In this talk, we propose a unified way to describe the various metaheuristic algorithms that focus on common and important search components of optimization algorithms: representation, objective function, constraints, initial solution(s), and variation operators. In addition to large-scale search space, the problem is characterized by its variable mixed design space, it is very expensive, and it has multiple blackbox objective functions. Hence, this unified methodology has been extended to advanced optimization approaches, such as surrogate-based, multi-objective, and parallel optimization.

Nav Mustafee
Centre for Simulation, Analytics and Modelling (CSAM), University of Exeter Business School, UK
N.Mustafee@exeter.ac.uk

Hybrid Modelling and Simulation for Transdisciplinary Science: Modelling & Simulation (M&S) techniques such as discrete-event simulation (DES), agent-based simulation (ABS), and system dynamics (SD) enable the development of computational models that aid decision-making. With the growing complexity of systems being modelled, numerous studies deploy hybrid simulation (HS), combining multiple M&S techniques like DES and ABS to provide improved insights into the evolving system dynamics at different levels of resolution. Distinct from HS, hybrid modelling (HM) extends the M&S discipline by combining theories, frameworks, techniques and established research approaches that have existed as extant knowledge within various disciplines, and applying multi-, inter- and trans-disciplinary solutions to practice. Within M&S, there is limited evidence of using conjoined methods for building HMs. Where a stream of such research does exist, the integration of approaches is mostly at a technical level. The talk will motivate the need for a transdisciplinarity-enabling framework for M&S that supports the collaboration of research efforts from multiple disciplines, allowing them to grow into transdisciplinary research.

Ralph E. Steuer
Sanford Family Distinguished Chair in Business, Department of Finance, Terry College of Business, University of Georgia, Athens, Georgia, USA
ralphsteuer@gmail.com

Yue Qi
Department of Financial Management, Business School, Nankai University, Tianjin, 300071, China
yorkche@nankai.edu.cn

Exploring the nonnegativeness of portfolio weight vectors of equality-constraint-only model series and implications for capital asset pricing models: Computing optimal sets has long been a topic in multiple-objective optimization. Despite substantial progress, there are still research limitations in the multiple-objective portfolio selection and optimization areas. Principally, the optimal set structure and public-domain software for even three objectives are barely available. Alternatively, researchers scrutinize equality-constraint-only models and analytically and fully resolve them. Within this context, this paper makes theoretical contributions to the literature. Specifically, we prove in theorems general existence of positive elements and negative elements for the optimal set of these analytical methods. Practically, we prove that there can exist a nonnegative subset of the optimal set. Consequently, the possible existence profoundly endorses these analytical methods, because researchers bypass mathematical programming, analytically resolve, and pinpoint some nonnegative optima. Moreover, we elucidate these analytical methods’ alignment with capital asset pricing models (CAPM). Furthermore, we generalize for k-objective models.

Mohamed Dia
Research Group in Operations, Analytics and Decision Sciences (RGinOADS)
School of Business Administration, Faculty of Management, Laurentian University, Sudbury, ON, Canada
mdia@laurentian.ca

Chair, AFROS’ AWG on Efficiency and Productivity Analysis

Data Envelopment Analysis (DEA) is one of the most prolific decision-making techniques of the past four decades. Since its inception paper (Charnes, Cooper and Rhodes, 1978), DEA has extensively been applied to measure the performance or relative efficiency of private and public organizations across almost all sectors. The benefits of using DEA to assess the performance or
efficiency of decision-making units (DMUs) lie in its capacity to determine a comprehensive measure of relative efficiency, from a set of multiple inputs and outputs, and to perform a benchmarking analysis of the DMUs. Multiple research reviews of theoretical and applied studies in DEA have been conducted and published in the recent years. In this talk, I will introduce the DEA methodolgy, its basic models, and an overview of its applications.