100PhDs - focus on Amal Machatalay
© 2025 EPFL
We are pleased to highlight the work of Amal Machatalay, supervised by Prof. Ahmed Ratnani (UM6P) and Prof. Daniel Kressner (EPFL), who proposes an innovative approach to modelling and simulating the complex dynamics of multi-class traffic. This research illustrates the potential of advanced numerical methods to address contemporary challenges related to mobility and the rise of autonomous vehicles.
By combining game theory, high-performance computing (HPC) methods and sophisticated numerical solution techniques, the researchers have extended a framework initially designed for homogeneous vehicles to realistic situations involving multiple vehicle types, including cars and trucks. Their work is distinguished by the development of a multi-class traffic model based on Mean-Field Games, capable of accurately reproducing strategic behaviour in a dense traffic environment.
Using regularised LGMRES solvers and an HPC infrastructure, the team successfully conducted very large-scale simulations involving up to 88 million variables. These advances not only provide a better understanding of the complex interactions between different vehicle classes, but also validate, through detailed comparisons between macroscopic and microscopic models, the relevance of the Nash approximation in these contexts, with results slightly better than theoretical expectations.
By providing robust digital tools for intelligent traffic modelling, this research perfectly illustrates how scientific innovation can support major technological transformations, particularly in the field of autonomous transport.
100PhDs for Africa is a programme from the Excellence for Africa initiative jointly led by Mohammed IV Polytechnic University in Morocco and EPFL.