Could you describe what it is you do ?
Atoptima is a decision-support software publisher that industrialises cutting-edge mathematical optimisation technology. We design and develop tailored optimisation modules to help with the planning of operations and vehicle routes, thus significantly enhancing the performance of systems.
Our solutions (which come in Software-As-A-Service form) are easy to integrate and scalable, while our optimisation solvers are capable of handling even the most specific constraints. What sets us apart is our global approach to handling complex resource-allocation decisions, which allows us to identify both the most appropriate and the most effective solution.
Atoptima is a deeptech made up of experts in operational research, with more than 25 years’ worth of experience in combinatory optimisation and a multidisciplinary team that is skilled in mathematical modelling, business intelligence, expert systems and web interfaces.
What are the technological challenges facing the sector ? And what's next for the start-up ?
Scaling up is the biggest technological obstacle for decision-support optimisation. The size and complexity of the problems encountered is leading to a rapid increase in the number of variables. Our solution has to become intelligent, developing approximations for organising searches in order to identify the right candidates, without the need for exhaustive lists. Because every advance means pushing the boundaries further, this is a permanent challenge: handling multiple decision levels simultaneously within an integrated approach, factoring in both the long-term and the short-term, and taking uncertainties into account in order to devise solutions capable of withstanding vagaries.
Because every advance means pushing the boundaries further, this is a permanent challenge.
Atoptima has two key challenges in this high-performance dynamic: to consider the generic programming of algorithms for intersecting use in applications, and to industrialise their implementation in high-level languages, ensuring they are both modular and maintainable. In this way, Atoptima delivers the best in DeepTech for the socio-economic sphere, with researchers combining the pursuit of excellence with a desire to tackle real market problems.
What did you do in your career before ? What motivated you to launch the start-up ? What support did Inria provide ?
Prior to founding Atoptima, I had an international academic career, studying at the Catholic University of Leuven (Belgium), Massachusetts Institute of Technology (USA), the University of Cambridge (UK) and the University of Bordeaux. I also took part in a number of industrial collaborations. In 2008 I set up RealOpt*, an Inria project team dealing with mathematical optimisation from which Atoptima emerged as a spin-off.
Atoptima came about through the technological maturation that took place within RealOpt, with support and assistance from Unitec. The Institute had the foresight to predict the use of high-performance global optimisation in handling combinatory complexity in operational planning. The opportunity to recruit the very best and to establish the scientific project long-term were both motivating factors in the decision to set up Atoptima. From the outset, the aim was to share expertise with the industrial world in relation to real-life trade problems.
Inria's philosophy, which centres around teamwork, the complementary nature of methodological and software developments, and promoting opportunities for industrial transfer, is key in creating the ideal territory for start-ups like Atoptima to emerge.
* RealOpt is a joint project team between Inria, CNRS, the University of Bordeaux and Bordeaux INP within the IMB (UMR 5251) and the Inria Bordeaux Sud-Ouest research centre.