Optimized decision-making
It was while writing her operational research thesis that Luce Brotcorne started out on this path, working on the topic of pricing. Until the 1980s, prices were primarily determined based on an estimate of historic demands, but the problem took on another dimension when consumers became active participants in making choices. Rail and air travel are good examples of these new kinds of behaviour. When buying tickets, travellers compare prices, travel time, connections, times and take advantage of miles, discount cards etc. This results in an array of new variables to factor in to estimate a fare. This is where mathematical programming comes in, to model the behaviour of two entities with different objectives – in this case consumers who want the best price and airlines whose activities must be profitable. “Known as hierarchical decision-making processes, these optimization problems now encompass a wide range of issues from logistics, to security and energy ”, adds Luce Brotcorne. Based on these models, specific high-performance algorithms are developed.
Supported by Inria’s efforts to promote cross-border teams, Luce Brotcorne chose colleagues from Centrale Lille and Université Libre de Bruxelles to form the central core of her current team to work on this innovative topic. In 2015, the Inocs team was created. Since then, the team has worked on a steady stream of projects. “We work with start-ups and major firms alike. And each project gives us the opportunity to conduct original research ”.
Anticipating change
The team leads research on theoretical problems in order to anticipate changes in society that companies will have to face. One such company is EDF , for which a thesis will be completed in 2018. Faced with the major changes the sector is undergoing – fluctuating demand, offshore production, intermittent generation and increased competition – it must redefine its system. “We’re also working with DHL on a project funded by the French National Research Agency to assess the future of the world of logistics where stock markets and auctions are increasingly present ”, says Luce Brotcorne. In terms of applications, Inocs has teamed up with companies for short-term collaboration to solve concrete problems. “Companies have specific constraints and we provide tailor-made solutions. ” It is also an opportunity to create value for local companies like Colisweb , the Lille-based delivery company. Inocs has developed computer modules for the start-up, making it possible to optimize deliveries based on order deadlines and availability of delivery contractors. The company Jules has also benefited from the team’s services through algorithms designed to facilitate logistics by determining where products should be stored in manual warehouses.
The group also has a strong international reputation. In addition to being a multicultural team, Inocs works with a number of international partners, two of which are part of the Inria associate teams program. The first is located in Chile, as part of the Inria Chile program, and works with the team on issues related to fraud prevention by modelling strategies for positioning patrols while taking into account those of the fraudsters who try to avoid them. The other is located in Montreal, and works on combining machine learning and artificial Intelligence approaches with optimization models developed by the team.
More information
Inocs is a joint project-team Inria, CRIStAL (Centrale Lille, CNRS, Université de Lille) and Université Libre de Bruxelles.
- Inocs - Integrated Optimization with Complex Structure
- Twitter @TeamINOCS
- Inria Chile
- Inria Chile on inria.fr