Dr. Thé Van LUONG - Lausanne, Switzerland
  • R&D project services, founder, development, scientific research and teaching.
  • Ph.D. in computer science at INRIA France and expertise on operations research.
  • Experience in solving industrial problems and designing dedicated optimization methods (metaheuristics and exact methods). Other interests focus on the parallelization of AI algorithms on CPU and GPU architectures.
  • Born in 1983 in Besancon (France).
  • thevanluong@gmail.com - Last update: 30 may 2021

Thesis: Metaheuristics on GPU


Having the opportunity to contribute among the first and pioneering GPU works during the CUDA 2.0 era, the goal of my thesis (2008-2011) was to redesign parallel models of metaheuristics on GPU architectures.

Ph.D. thesis: Parallel Metaheuristics on GPU.
[ .bib ] [ .pdf ] (slides:pdf)

Free access to my articles here. Citations available in DBLP. Google "metaheuristics GPU".

With the proliferation of GPU works, the parallelization of simple algorithms on GPU should be nowadays accessible to any scientist.

Services and projects


> R&D projects on planning and scheduling services. We provide dedicated solutions for enterprises and institutions.

> Founder of an Ethereum project on GPU:
The Ojomine project - 51 machines in commercial premises in Grandson.

> Transports publics genevois (TPG - HEIG-VD): R&D project management and expertise, design algorithms for the automated generation of parking plans, application development, maintenance and customer service. French demonstration here .



Solving industrial problems


Industrial problems: unprecedented size, complex modelling (operational constraints, complex data structures and undefined objective functions) and limited execution time.

For example, to design a parking plan of hundreds vehicles for a transportation company where each vehicle can be positioned in 300+ locations. Traditional approaches might fail to provide a solution compatible with production requirements.

It is very interesting to see the tremendous gap between dedicated solutions and commercial solvers for solving industrial problems.