How coordination can boost the resilience of complex supply chains

By Célian Colon, PhD student at the Ecole Polytechnique in France & IIASA Mikhalevich award winner

How can we best tackle risks in our complex and interconnected economies? With globalization and information technologies, people and processes are increasingly interdependent. Great ideas and innovations can spread like wildfire. However, so can turbulence and crises. The propagation of risks is a key concern for policymakers and business leaders. Take the example of production disruption: with global supply chains, local disasters or man-made accidents can propagate from one place to another, and generate significant impact. How can this be prevented?

Risk propagation is like a domino effect. Credit: Martin Fisch (cc) via Flickr

Risk propagation is like a domino effect. Credit: Martin Fisch (cc) via Flickr

As part of my PhD research, I met professionals on the ground and realized that supply risk propagation is a particularly tricky issue, since most parts of the chains are out of their control. Supply chains can be very long, and change with time. It is difficult to keep track of who is working with whom, and who is exposed to which hazard. How then can individual decisions mitigate systemic risks? This question directly connects to the deep nature of systemic problems: everyone is in the same boat, shaping it and interacting with each other, but no one is individually able to steer it. Surprising phenomena can emerge from such interactions, wonderfully illustrated by bird flocking and fish schooling.

As an applied mathematician thrilled by such complexities, I was enthusiastic to work on this question. I built a model where firms produce and interact through supply chain relationships. Pen and paper analyses helped me understand how a few firms could optimally react to disruptions. But to study the behavior of truly complex chains, I needed the calculation power of computers. I programmed networks involving a large number of firms, and I analyzed how localized failures spread throughout these networks, and generate aggregate losses. Given the supply strategy adopted by each firm, how could systemic risk be mitigated?

With my collaborators at IIASA, Åke Brännström, Elena Rovenskaya, and Ulf Dieckmann, we have highlighted the key role of coordination in managing risks. Each individual firm affects how risks propagate along the chain. If they all solely focus on maximizing their own profit, significant amounts of risk remain. But if they cooperate, and take into account the impact of their decisions on the risk profile of their trade partners, risk can be effectively mitigated. Reducing systemic risks can thus be seen as a common good: costs are heterogeneously borne by firms while benefits are shared. Interestingly, even in long supply chains, most systemic risks can be mitigated if firms only cooperate with only one or two partners. By facilitating coordination along critical supply chains, policy-makers can therefore contribute to the reduction of risk propagation.

Colon's model analyzes how firms produce and interact through supply chain relationships. Credit: Jan Buchholtz (cc) via Flickr

Colon’s model analyzes how firms produce and interact through supply chain relationships. Credit: Jan Buchholtz (cc) via Flickr

Drawing robust conclusions from such models is a real challenge. On this matter, I benefited from the experience of my IIASA supervisors. Their scientific intuitions greatly helped me focusing on the most fertile ground. It was particularly exciting to borrow techniques from evolutionary ecology and apply them to an economic context. Conceptually, how economic agents co-adapt and influence each other shares many similarities with the co-evolution of individuals in an ecological environment. To address such complex issues, I strongly believe in the plurality of approaches: by illuminating a problem from different angles, we can hope to see it more clearly!

Note: This article gives the views of the author, and not the position of the Nexus blog, nor of the International Institute for Applied Systems Analysis.

The mathematics of love

By Sergio Rinaldi, IIASA Evolution and Ecology Program and Politecnico di Milano, Italy

Is it possible to predict how love stories develop, progress, and end using mathematical models? I have studied this question over the past 20 years with a group of researchers at IIASA and at the Politecnico di Milano, and as we show in our new book Modeling Love Dynamics (World Scientific, 2016), the answer is yes. The emerging message is that prediction is possible, if we can describe in formulas the way each individual reacts to the love and to the appeal of the partner.

Consider a standard love story, which develops like those described in a classical Hollywood movie such as Titanic. This story can be easily modeled, if one considers reasonably appealing individuals who increase their reaction with the partner’s love – so called secure individuals. Starting from the state of indifference, where the individuals are at their first encounter, their feelings continuously grow and tend toward a positive plateau.

Mala Powers & José Ferrer in Cyrano de Bergerac, 1950. Credit: Public Domain

Mala Powers and  José Ferrer in Cyrano de Bergerac, 1950. – Public Domain

Love stories become more intriguing when one individual is not particularly appealing, if not repelling, as in the fairy tale “Beauty and The Beast.” Indeed, in these cases, there exists also a second romantic regime, which is negative and can therefore entrain, in the long run, marital dissolution. In order to avoid that trap, people who are not very charming, or believe to be so, do all they can to look more attractive to the partner. At the first date, she wears her nicest dress and he shows up with his best fitting T-shirt. However, after a while, the bluffing can be interrupted, because the couple has entered the safe basin of attraction of the positive regime. Needless to say, the model also supports much more sophisticated behavioral strategies, like that described by Edmond Rostand in his “Cyrano de Bergerac,” the masterpiece of the French love literature.

Not all individuals are secure. Indeed, some people react less and less strongly when the love of the partner overcomes a certain threshold. These individuals, often very keen to flirtation, are incapable of becoming one with their partner. The model shows that couples composed of insecure individuals tend, with almost no exception, toward an unbalanced romantic regime in which the most insecure is only marginally involved and is therefore prone to break up the relationship at the first opportunity. This is why after just 20 minutes of the very long “Gone with the Wind,” when one realizes that Scarlett and Rhett are both insecure, the model can already predict the end of the film, where he quits her with the lapidary “Frankly, my dear, I don’t give a damn.” The same conclusion is expected if only one of the two individuals is insecure. This explains the numerous failures in the romantic life of some individuals, like the beautiful star Liz Taylor, who is described as very insecure in all her biographies, and went, indeed, through eight marriages.

Clark Gable and Vivien Leigh in "Gone with the Wind."

Clark Gable and Vivien Leigh in Gone with the Wind, 1939 – MGM Pictures | Public Domain

Mathematical models can also be used to interpret more complex romantic behaviors. Particularly important is the case of individuals who overestimate the appeal of the partners when they are more in love with them (like parents who have a biased view of the beauty of their own kids). Interestingly, if insecurity is also present, biased couples can have romantic regimes characterized by recurrent ups and downs. In other words, the theory says that bias and insecurity is an explosive mix that triggers turbulence in the life of a couple.

In the second part of the book we focus on  the effects of the social environment and to the consequences of extra-emotional compartments. In this context, our analysis of the 20-years long relationship between Laura and the famous Italian poet Francis Petrarch shows that poetic inspiration is an important destabilizing factor, responsible for transforming a quiet relationship into a turbulent one.

Finally, we studied triangular relationships, with emphasis on the effects of conflict and jealousy. In all these cases the dynamics of the feelings can be very wild, up to the point of being chaotic and, hence, unpredictable. When this occurs, the life of the couple becomes unsustainable, because painful periods of crisis can virtually start at any moment: a heavy permanent stress. The model can thus explain why the relationship is often interrupted, sometimes even tragically, as in the famous film by François Truffaut “Jules et Jim”, where Kathe’s suicide is perceived as a real relief.

CaptureMore information: Watch a video of Sergio Rinaldi’s talk at the 2015 Systems Analysis Conference.

Reference
Rinaldi S, Della Rossa F, Dercole F, Gragnani A, Landi P, (2015). Modeling Love Dynamics. World Scientific, Singapore [January 2016]  http://www.worldscientific.com/worldscibooks/10.1142/9656

 

Note: This article gives the views of the author, and not the position of the Nexus blog, nor of the International Institute for Applied Systems Analysis.

Don’t dope – run the code!

By Andrey Krasovskii, IIASA Ecosystems Services and Management Program

During his workout in the IIASA gym, my colleague Pekka Lauri often runs on a treadmill. He adjusts the velocity of running using the control panel, and it indicates the distance and approximate calories burnt. While Pekka may not be thinking about mathematical models during his workout break, running and other athletic performance can be modeled using some of the same techniques that we use for other questions at IIASA.

Outside of my academic work at IIASA, I am highly interested in sports, and athletics in particular. My wife, Katy Kuntsevich, has won Austrian championships in high jump several times, and my brother, Nikolay, was on his university team as a 400 meter runner.

Visiting my hometown Ekaterinburg, Russia, back in 2014, I got into a dispute with my father, who is a university professor in theoretical mechanics. Namely, my argument was that the sprinters’ acceleration at the finish line of the 100 meter race should be negative. Later on, during the Christmas holidays, I decided to mathematically support my statement.

Left: Prof. Krasovskii with a page from his Lectures in Theoretical Mechanics (Chapter 2, Kinematics, in Russian) featuring Valeriy Borzov on finish. This page was a reason for my study. Top right: Portrait of Isaac Newton.

Left: Prof. Krasovskii with a page from his Lectures in Theoretical Mechanics (Chapter 2, Kinematics, in Russian) featuring Valeriy Borzov on finish. This page was a reason for my study. Top right: Portrait of Isaac Newton.

When athletes run a race, their horizontal velocity can be estimated by modern technologies such as high resolution cameras and lasers. Knowing the horizontal instantaneous velocity, one can calculate acceleration. According to Newton’s law, one can introduce the forces applied to the body center mass of the athlete. Outside a gym treadmill it gets a bit more complicated, with air resistance and headwind or tail-wind. Against these aerodynamic forces the runner applies horizontal force, which drags him forward. In reality it is an “aggregate” of impulsive normal forces generated by feet and the stroke frequency.

The dynamics of an athlete have been described by an ordinary differential equation studied in papers on sprint modeling, first published by physicist J. Keller. This equation has been considered in numerous papers, which have shown that the equation fits the real data for short-distance running (100 and 200 meters). The indicated studies are devoted to the calibration of parameters, to the wind impact analysis, and to the choice of force functions such that the solution satisfies the actual motions, e.g. the running records of Usain Bolt.

The problem of running dynamics reminded me of a type of model that we sometimes use at IIASA, called an optimal control model. Optimal control models are used to calculate the best or most efficient way of doing something, for example, driving from one place to another. If I want to drive from Vienna to Laxenburg, I start my car’s engine at my house (point A). I look at the time, and plan to arrive to IIASA (point B) in 30 minutes. In optimal control terms, the car is a control object, and the driver controls it by pushing the gas/brake pedals and steering the wheel. In the driving process the car meets certain constraints (e.g. the engine power, available roads, and speed limits) and disturbances (e.g. traffic jams, lights, and weather conditions). Obviously, there are many ways of controlling the car in order to reach IIASA in 30 minutes. What if among those admissible controls, I wanted to find an optimal control minimizing car’s energy expenditures during the 30-minute trip from A to B? Here energy is an intensity (cost) of control actions, i.e. fuel (petrol/electricity), or corresponding greenhouse gas emissions. Well, in this case one needs to solve the classical minimum energy control problem. The solution to this problem gives an optimal plan that the car driver (or autopilot) needs to implement. Note, that the corresponding time-optimal control problem consists in finding the fastest driving time to IIASA under given fuel reserve. Optimal control theory (OCT) is an efficient tool for solving such dynamic optimization problems.

My research question was: “Can one control his/her running similar to driving a car?” The answer is: “Yes!”

I applied an optimal control model to Usain Bolt’s performance data at the Beijing Olympic Games in 2008, when he ran the 100 meter sprint in 9.69 seconds. According to the model, under the same conditions he could have distributed his energy optimally and run the distance in 9.56 s. It is worth mentioning that this time is close to his current world record, 9.58 s, achieved at the 2009 World Championships in Berlin. In the paper I also provide modeling results for optimal (energy-efficient) running over 100 m: calculation of the minimum energy and trajectories of acceleration, velocity, and distance from start.

In the conclusion, I argue that applying advanced science in the athletic training programs is far better than doping–better in terms of a healthy body, mind and soul.

Here is my hypothesis of what Usain Bolt is doing at his laptop. © Weltklasse Zürich - Marcel Giger

Here is my hypothesis of what Usain Bolt is doing at his laptop. © Weltklasse Zürich – Marcel Giger

Reference:
A. A. Krasovskii, “Application of optimal control to a biomechanics model”, Proceedings of the Steklov Institute of Mathematics, 2015, Vol. 291, pp. 118–126. http://dx.doi.org/10.1134/S0081543815080118

 

I would like to thank Sergey Aseev, Katherine Leitzell, as well as my colleagues in the IIASA Ecosystem Services and Management Program (ESM) for their interest and valuable discussions.

 

Note: This article gives the views of the author, and not the position of the Nexus blog, nor of the International Institute for Applied Systems Analysis.

This summer in Moscow: Impressions from Moscow Summer Academy 2015

By M. Nazli Koseoglu, MSA 2015 Participant, School of Geosciences, Environmental Economics, Edinburgh University, Scotland

M. Nazli Koseoglu

M. Nazli Koseoglu

The Summer Academy on Economic Growth and Governance of Natural Resources took place at Lomonosov Moscow State University from 20th July to 1st August 2015.

As an environmental economist working on economic valuation and optimisation of water use, the academy was very interesting for me. Water management is a dynamic process and requires bringing perspectives and expertise from different disciplines together. Application of systems analysis enables us to combine aspects from various domains, come up with models that identify nonlinearities, project regime shifts, and tipping points in the management of water as well as other natural resources. Such projects require interdisciplinary collaboration and communicable results to inform policy. Scientists need to translate their results to a language accessible to the policymakers, in order for society to pick up on and capitalize on the research efforts. The MSA 2015 provided me with necessary training to go deeper into different modelling methodologies, and learn the concepts and principles of science for policy first-hand from IIASA scientists.

The reading list sent before the course gave me the impression that I would probably be the only environmental economist amongst a crowd of mathematical modellers. However, arriving in Moscow, I found that the MSA 2015 participants came from a broad range of backgrounds and countries at different stages of their careers in academia or policy. We all came  to learn and discuss the natural resource constraints to infinite economic growth on finite planet.

During lectures, the theoretical foundations of different mathematical approaches such as dynamical systems theory, optimal control theory and game theory were presented by leading scientists, such as Michael Ghil. Fundamentals of addressing challenges of natural resource management and comparing contemporary models of economic growth were also covered as central themes.

The course acknowledged the issues related with ecosystems services, public goods, inter-generational and international fairness, and public and common pool resource dynamics in the face of economic growth and resource constraints. The training underlined feedbacks between institutional dynamics and resource dynamics in complex social-ecological system and need for interdisciplinary and policy-relevant research, an important take-home message for next generation scientists.

Photos by M. Nazli Koseoglu

Photos by M. Nazli Koseoglu

What makes the MSA so special?
Apart from lectures, we had tutorials, a group project, poster and project presentation sessions, as well as interesting talks on IIASA activities by Margaret Goud-Collins and Elena Rovenskaya, and an inspiring session on the importance of finding the right mentor for a successful career by Prof Nøstbakken. The  MSA 2015 program had a good balance of theory and practice, which encouraged participants to be proactive and engaged.

I particularly liked the poster session. We presented our ongoing projects and received feedback from the lecturers and other participants. It was great to get comments and perspectives that I never thought of, and tips from senior researchers. In the late days of the academy we were assigned to prepare a group project on Artic systems which allowed us to put what we had learned at the lectures into practice and apply important topics outside our exact fields of study; in my case, these topics were petroleum economics and artic futures. I found the multi-disciplinary group work to be a great exercise for the development of my current study.

Attending the MSA 2015 provided useful training, both theoretical and practical, for understanding systems analysis approaches better. The host institution and organizing committee at Lomonosov Moscow State Univesrity provided impeccable hospitality, and the setting, in a landmark building in a landmark city, was a great perk. I received very constructive feedback, and made good connections around the world. I would recommend all early-career researchers in relevant fields to take this great opportunity next summer!

More information about MSA 2015

Participants in the MSA 2015. Photo Credit: M. Nazli Koseoglu

Participants in the MSA 2015. Photo Credit: M. Nazli Koseoglu

Note: This article gives the views of the author, and not the position of the Nexus blog, nor of the International Institute for Applied

Systems Analysis.

 

10 steps to removing carbon from the global economy

By Nebojsa Nakicenovic, Deputy Director General, International Institute for Applied Systems Analysis (IIASA), Austria (Originally published on the World Economic Forum Agenda Blog.)

Nebojsa Nakicenovic

Nebojsa Nakicenovic

Goal 7 of the Sustainable Development Goals is ambitious: Ensure access to affordable, reliable, sustainable and modern energy for all. This must be accomplished without compromising Goal 13: climate. This is achievable.

In spite of ups-and-downs and outright shocks in the global economy, some quite recent, the economic success stories of the industrialized countries are role models for the countries that are still developing. This puts the entire global community in the dichotomous position of needing to fire up the engine of growth, without producing the greenhouse gases it has been emitting since the beginning of the Industrial Revolution. What is the answer?

Very few questions in the complex area of energy and climate change can have a simplistic answer, but I am going to attempt one here: decarbonization, namely, drastic reduction of carbon dioxide and other greenhouse gas emissions per unit of economic activity.

Back in 1993, I wrote this:

“The possibility of less carbon-intensive and even carbon-free energy as major sources of energy during the next century is consistent with the long-term dynamic transformation and structural change of the energy system.”

My view in 2015 is the same; however, the scientific community 22 years later has a much better understanding of “the decarbonization challenge” and how it can be addressed. I will sketch out a 10-step approach to the removal of carbon from the global economy, but first I’d like to paint in a bit of the background.

Carbon dioxide is the main greenhouse gas and contributor to climate change. The largest source is our use of fossil fuels to drive development. Carbon dioxide emissions have increased exponentially since 1850 at about 2% per year, while decarbonization of the global economy is only around 0.3% per year.

The 2012 Global Energy Assessment, in which IIASA played a leading role, puts the current decarbonization rate at approximately six times too low to offset the increase in global energy use of about 2% per year. To meet the goal of the 2009 climate agreement (the Copenhagen Accord), namely, “the scientific view that the increase in global temperature should be below 2 degrees Celsius” to prevent dangerous anthropogenic interference with the climate system, global net emissions of carbon dioxide and other greenhouse gases will need to approach zero by the second half of this century, implying deep, deep decarbonization rates.

working oil pumps © Kokhanchikov | Dollar Photo Club

“Carbon dioxide is the main greenhouse gas and contributor to climate change. The largest source is our use of fossil fuels to drive development.” © Kokhanchikov | Dollar Photo Club

But we need deep decarbonization while energy needs are increasing to meet the demand of the developing world, including the three billion without access today to sustainable energy. All scenarios in the academic literature that lead to further economic development in the world, universal access to sustainable energy, and the stabilization of climate change to less than 2 degrees Celsius, anticipate deep and urgent decarbonization. Here’s my 10-point plan for doing that.

  1. Change attitudes
    Attitudes to energy use are based on many factors, from cultural norms to overall infrastructure design. We need much greater political will to affect a change in attitudes: it is critical that policy interventions should communicate to citizens the ethical notion of improved well-being and health now and for future generations of a zero-carbon economy. .
  1. Transform governance
    The transformation needed this century is more fundamental than previous transformations, like the replacement of coal by oil, because of the significantly shorter time needed to achieve it. Thus, government policies are essential, and are needed particularly in changing buildings codes, fuel efficiency standards for transportation, mandates for the introduction of renewables, and carbon pricing.
  1. Improve energy efficiency
    More efficient provision of energy services, or doing more with less, and radical improvements in energy efficiency, especially in end use, will reduce the amount of primary energy required and represents a cost-effective, near-term option for reducing carbon dioxide emissions, as well as having multiple benefits in different areas of life.
  1. Ramp up renewable use
    We can show that the share of renewable non-fossil energy from solar, wind, rain, tides, waves, and geothermal sources in global primary energy could increase from the current 17% to between 30% and 75%. In some regions it could exceed 90% by 2050, provided that public attitudes change and efficiency increases.
  1. Reduce global energy intensity
    The energy intensity in the industrial sector in different countries is steadily declining due to improvements in energy efficiency and a change in the structure of the industrial output. Far greater reductions are feasible by combining these improvements with adoption of the best-achievable technology.
  1. Use known technologies
    Carbon dioxide capture and storage (CCS), now being piloted, is a pathway that leads to decarbonization with continued use of fossil energy. It requires: reducing costs, supporting scale-up, assuring carbon storage integrity and environmental capability, and securing approval of storage sites. Nuclear energy could make a significant contribution in some parts of the world, or it could be phased out as, for instance, in Germany.
  1. Improve buildings
    Retrofitting buildings can reduce heating and cooling energy requirements by 50–90%; new buildings can be designed and built using close to zero energy for heating and cooling. Passive energy houses and those that produce energy onsite are another great opportunity to achieve vigorous decarbonization. In conjunction with compatible lifestyles oriented toward rational energy use, efficient buildings are an important decarbonization option.
  1. Cut transport carbon
    A major transformation of transportation is possible over the next 30–40 years and will require improving vehicle designs, infrastructure, fuels and behavior. Electrically powered transportation reduces final energy use by more than a factor of three over gasoline-powered vehicles. A shift toward collective mobility is an essential option. This also implies behavioral changes and new business models like car-sharing, and self-driving cars to replace individual mobility.
  1. Clean industrial processes
    Overall, global industry efficiency is only 30%. Improved energy efficiency in industry results in significant energy productivity gains and, in turn, improved productivity boosts employment and corporate competitiveness. A shift toward low to zero emission energy sources in industry is another important and much-needed change. For example, with an aggressive renewables strategy, near-zero growth in GHG emissions in the industrial sector would be possible. Finally, decarbonization would also involve changes of industrial processes, for example, from high to low temperatures.
  1. Stranded assets and ‘derisking’ renewables.
    The flow of investment needs to be changed away from fossil fuels and toward efficiency, renewables, decarbonization of fossil energy sources, and especially efficient end-use in buildings, transport, and industry. Sustainable energy futures require relatively high up-front investments with the benefit of low long-term costs. They are attractive in the long run, but the up-front investments need derisking and other forms of support, such as feed-in tariffs.

Note: This article gives the views of the author, and not the position of the Nexus blog, nor of the International Institute for Applied Systems Analysis.