13 January 2014

# A passion for science

December 2013. Options Magazine, Winter 2013/2014
IIASA’s Elena Rovenskaya explains how IIASA researchers develop the models they use to understand complex systems

Elena Rovenskaya

Q Dr. Elena Rovenskaya, what’s a mathematical model?
A A simplification of reality, elegantly expressed through equations.

Q An oversimplification, perhaps?
A No, what we model is usually dictated by researchers’ interests or project requirements. We capture as much detail as needed. With today’s computing power, that can be a lot.

Q So you write equations…
A Yes, sometimes one; sometimes 10,000.

Q What could one equation show?
A It could capture a conceptual process of economic development comprising different levels of authorities, thousands of companies, and millions of people—all within a standard framework of economic growth theory.

Q Give a recent example of how IIASA created a new model.
A We worked with a team from Finland’s Ministry of Economy, a country with an IIASA NMO. This comprised the Minister’s Advisor and representatives of three Finnish regions. We asked
about the regions’ current and past economic situations, the government’s goals, and what policies they had in mind for achieving them. As always in an open and friendly meeting,
our ideas became gradually “interconnected,” which made writing a good model easier.

Q Do you extrapolate past trends to project the future?
A Sometimes, but the future is going to see big structural changes in all global sectors. So we can’t use the past in a mechanical way. Old data are useful for “soft” validation purposes.
We can fit our model to data from, say, 1950–1970 to project outcomes for 1970–1990, and knowing what happened in the 70s and 80s, we can check our model is on track.

Q IIASA is building some big integrated models like GLOBIOM, which assesses the competition for land use between agriculture, bioenergy, forestry, and livestock. These have several modules—submodels. How do researchers “bolt” submodels together?
A Technically the models are tied together by a special code. Conceptually, you need to turn outputs from one submodel into inputs to another. This may change the performance of each submodel; so again it’s not mechanistic, it requires brainpower!

Q Your modeling results don’t always agree with those of other institutions.
A
Each model looks at reality from its own angle. So it’s natural for results not to coincide.

Q What does integrated modeling show you?
A
It gives a tremendous overview of the problems, as well as profound insights into the co-benefits and trade-offs needed to achieve the required goals.

Q What’s good about modeling at IIASA?
A
The freedom to explore and experiment. There aren’t he constraints you’d find in academia. Maybe not all our studies are successful, but there’s a chance to make
enormous breakthroughs.

Q And finally, why do Russians have a reputation for being so good at math?
A
Our teachers in Russia came from a generation ith a passion for science. It was reflected everywhere, for example in science fiction. Scientists, especially physicists and mathematicians, were heroes in life and in art, and to us they had the most interesting lives. This was a very strong—and romantic—motivation to become a good scientist.

Elena Rovenskaya

Acting Leader of IIASA's Advanced Systems Analysis Program and Research Scholar at the Optimal Control Department of the Faculty of Computational Mathematics and Cybernetics, Lomonosov Moscow State University

Interview by Kathryn Platzer

Last edited: 09 October 2014

CONTACT DETAILS

Elena Rovenskaya

Acting Deputy Director General/CEO

Directorate - DDG For Science

Program Director