A new research group in Paris aims to use optimisation theory to find answers to the challenges of ensuring an efficient and sustainable future.
How do you improve an electric grid that was not conceived for electric vehicles and renewable energy? How to best manage natural resources, whether fossil fuels or biodiversity? Could software for managing laptop batteries be used to control temperatures in a building?
These are complex questions, and they and many others may find their answers in a branch of applied mathematics known as optimisation or mathematical programming. And because optimisation produces decision-making tools, no issue may need more urgently support from this science than sustainable development.
That is why Microsoft, the École Polytechnique of Paris and the Institute of Information and Engineering Sciences and Technologies of the CNRS, France’s national research body, have just created a new chaire – research team – in Optimisation and Sustainable Development. The team is setting out to adapt or discover algorithms that will improve industrial, economical and ecological efficiency and decision-making in complex fields such as energy, natural resources, transport, and urban or construction planning.
‘Because the variables are often fuzzy, it requires a lot of discussions with experts to build a model that includes all the constraints of the real world.’ - Philippe Baptiste, director of the joint computer sciences laboratory of the École Polytechnique and the CNRS
“Optimisation has its roots in the period following World War II and has been used in some industry since the 1960s,” explains Youssef Hamadi, the Constraint Reasoning Group’s leader at Microsoft Research in Cambridge, UK, who will co-head the new chaire. “It is, for example, currently used in transport to improve the distribution routes of a fleet of trucks.”
With optimisation, researchers build mathematics models and tools that will maximise or minimise a real function of any given system by fine-tuning its underlying variables. “The difficulty in complex systems is to build those models on the basis of data that are by nature uncertain and dynamic,” adds Hamadi.
“So we use stochastic programming – probabilities – to elaborate those models.”
“Because the variables are often fuzzy, it requires a lot of discussions with experts to build a model that includes all the constraints of the real world,” explains Philippe Baptiste, the other co-head of chaire and director of the joint computer sciences laboratory of the École Polytechnique and the CNRS.
“During the past 15 years, we have gathered more and more data to build bigger and bigger models. One solution to handle those models was to increase the processing power of computers. But it has proved insufficient. Mathematics and more specifically optimisation appear now the best way to handle those large set of complex data,” says Baptiste.
That is particularly true in sustainable development, where not only are data fuzzy but objectives may be multiple, economical and ecological, multi-scaled (with criteria such as the total amount of energy used, and the corresponding carbon footprint) and embed concepts of equity.
Optimisation has already helped to choose the best migration corridors for protected animals. It helps planners to choose where to site buildings and infrastructures that will minimise the impact of human activities on the wildlife. It is also an important tool in choosing where to put sites such as data centres, which are heavy users of water and electricity. Baptiste also thinks that optimisation will play a determining role in the redesign of airlines routes for an industry under pressure from carbon limits and the price of oil.
“The good thing about optimisation is that it forces you to think deeply about what you do,” says Baptiste. “In sustainable development you also have to think about the long-term impact of an activity; with such complexity and with so many variables optimisation is the only hope of finding solutions.”
The new initiative is in the process of hiring ten researchers to apply optimisation to concrete issues of sustainability faced by industrial companies. It may benefit from the input of the five chaires in the field of sustainable development that large companies, such as cement maker Lafarge and electric utility EDF, have established on the campus of the École Polytechnique.
Being based at the heart of the emerging unified Saclay Campus, to the southwest of Paris, the chaire may receive other inputs from companies associated in the Digiteo research cluster. The day the new chaire was inaugurated, representatives of companies such as electric components makers Alstom and Schneider were presenting their needs in terms of optimisation to make their business more sustainable.
“There is a lot of emerging activities where optimisation could play a critical role to improve things,” Baptiste believes. “Take the example of an electric vehicle. Tomorrow an algorithm may be able to manage the electricity it has in its battery. It could choose to use it for home appliances when the car is plugged in the garage because it knows, from the habit of the owner, that he does not use his car at night and that grid electricity is more expensive now than later, when it will reload to have its batteries full in the next morning.”
“A similar mathematical tool can help you manage the fish stock in a given zone of the Atlantic or the Mediterranean to make the resource more sustainable,” adds Hamadi.
Optimisation is all about making things better. With sustainable development sharing the same objectives, the marriage of mathematics and ecology created by Microsoft, École Polytechnique and CNRS is not only logical, but has a high probability of becoming fruitful.