Tag Archives: Global Systems Science

Dependently-typed programming in scientific computing: Examples from economic modelling

Cezar Ionescu (at PIK) and Patrik Jansson (me, at Chalmers) have just got a paper accepted which fits in well in the GSS activity.

Pre-print + abstract


Computer simulations are essential in virtually every scientific discipline, even more so in those such as economics or climate change where the ability to make laboratory experiments is limited. Therefore, it is important to ensure that the models are implemented correctly, that they can be re-implemented and that the results can be reproduced. Typically, though, the models are described by a mixture of prose and mathematics which is insufficient for these purposes. We argue that using dependent types allows us to gradually reduce the gap between the mathematical description and the implementation, and we give examples from economic modelling. We discuss the consequences that our incremental approach has on programming style and the requirements it imposes on the dependently-typed programming languages used.

Global Systems Science – II

At its second meeting the group had about twenty people. The intention was to discuss the document “Towards a global systems science” by Ralph Dum.
The was considerable discussion of the first paragraph: “Global Systems Science (GSS) is a response to two major 21st century developments, one societal and one technological: The increasingly global and highly interconnected nature of challenges facing humanity and the pervasiveness of Information and Communication Technologies – ICT – in all human and societal endeavours.”
It was noted that the study of social systems is not neutral. Different paradigms imply different models.
Why was ICT included? There were two answers to this question, one scientific and the other administrative. It was noted that the document highlights the term “policy informatics” – building ICT systems to support policy. All organisations use computer tools to support their planning and decision making (the tools may not be very good and the scientific principles they embody may at variance with observation). It was stated that the science we do is ICT-based and that our science is entangled with ICT. It was suggested that “societal informatics” could mean the use of computers to investigate social systems and social processes. This can be curiosity-driven science. It can augment policy informatics by providing science knowledge when it does not exists. Another category, ‘embedded informatics’ was suggested to reflect the fact that our societies have informatics embedded throughout, often supplied and supported by commercial organisations. It is necessary to configure data to answer questions. It was noted that ICT contributes to the “highly connected” nature of societies.
The second reason for considering ICT is the relationship between the scientific community and the European Commission. The complex systems community in Europe has been well supported by the EC, especially the Future Emerging Technology (FET) unit “FET is the ICT incubator and pathfinder for new ideas and themes for long-term research in the area of information and communication technologies. Its mission is to promote high risk research, offset by potential breakthrough with high technological or societal impact.” (http://cordis.europa.eu/fp7/ict/programme/fet_en.html). This meeting is intended to make the concept of ‘Global Systems Science’ well defined and for the research community to reach consensus and have a shared vision of the future. This is important because the Commission is developing its Horizon 2020, the next Framework Programme for Research and Innovation which is the successor to FP7. It will be launched in 416 days from 10/11/12 (http://ec.europa.eu/research/horizon2020/index_en.cfm?pg=home&video=none). By making Global Systems Science well defined, tangible, and obviously relevant to the Commission’s objectives the research community can ensure that there will be funding streams for GSS in H2020.
There was some discussion of what ‘complex systems’ means. For some it meant unexpected emergence. Clearly the suggested example of financial crisis, climate change, and urban dynamics are complex systems. The interpretation of ‘global’ was again discussed. Global can include ‘worldwide, as in climate change, but need not, as in cities. It was suggested that the term means the whole system with all its entangled subsystems, e.g. in cities the police, fire, education, retail, transport, health, etc subsystems are all interdependent. Policy requires that the whole system is considered.
Is it necessary to have integrated multilayer models? It was noted that there is the € 22m DYM-CS project addressing exactly this (http://cordis.europa.eu/fp7/ict/fet-proactive/dymcs_en.html) . Is the idea of networks of networks relevant or central to GSS. Some thought it definitely is, but this is only part of the story. The theory of networks has to change.

The term ‘Evidence Based Policy’ was discussed. Some people felt uncomfortable with this because often we don’t have much ‘scientific’ evidence. In the UK the policy makers use the term ‘Evidence Based Policy’ to give their policies legitimacy. However there are examples of the same scientific evidence being used to support contradictory policies.
It was agreed that there should be explicit exemplars of Global Systems.
Prof Wanglin YAN showed us the Global Environment Systems Leaders Programme at Keoi University in Tokyo. This resonates with GSS.
We need to establish legitimacy. Physicists have not been very successful in trying to displace (incorrect) ideas in economics. We need to educate many people.
The idea of ICT as an instrument for observation was discussed – ICT provides ’sensors’ for observing and measuring society. ICT isthe next telescope or microscope? CERN was mentioned. However, new types of ICT are required.
Although we did not go through Raph Dum’s document line by line, there was consensus that it is a very good start in the process of making Global Systems Science well defined and an idea that is useful for the research community

Much more was said. Sorry if your contribution got lost. If so please add it to the blog. JJ

Global Systems Science – Workshop I

The group included scientists and (at least) two people with experience of applying scientific methods in a policy context, one from a company with 300,000 employees worldwide.

The context for the discussion was the paper on GSS by Ralph Dum (available on this site).

There was consensus that Global System Science is policy oriented.

The term ‘global’ is interpreted as meaning the whole system and can apply at different levels including the city level, national level and international level worldwide. Global system have entangled subsystems that cannot be analysed in isolation, e.g. water, waste, power, transport, housing, crime, employment, climate, etc. Geography and history (path dependence) are usually important. Global systems have many levels.

There was (implicit) consesus that GSS involves systems of systems and networks of networks.

Some systems have to be modelled bottom-up from individual agents.

Although we believe GSS must involve policy, scientists are not good at interfacing to policy makers. The example was given of working with the mayor of a large city. These are smart people who handle complexity all the time. The example included talking to the mayor to understand how they see their problems, and pointing out that the issues are coupled and may be best handled from a systems perspective.

Global systems science can be defined around POLICY INFORMATICS, i.e. building ICT systems to address particular policy issues. The example was given of the $180 million tem year research programme at Los Alamos that started with the US Federal Clean Air Act which could not be implemented because no-one understood how road traffic caused polutants (it can be noted that this generated new science).

There was a discussion of computability.

There was (implicit) consensus that GSS involves Big Data.

Global System Science can be viewed in terms of coordination failure. This can include faliyre cascades. This was considered to be a fruitful direction.

Much more was discussed. The following points were made in conclusion:

* we need to make a bridge between scientists and policy makers
scientists need to usnderstand the language and methods of policy makers
scientists need to be better communicators (some are better at this than others)

* it is useful to consider coordination faliures and cascades of failure
we need a list of exemplar Global Systems
this can help to make the bridge

* we are close to having useful modelling of heterogeneous agents on (changing) (multilevel) networks (of networks)

+ there is the issue of computatiblity (related to topology, local v global) and new ideas in computation.

sorry for leaving a lot out, more to come …

In the meantime we all read Ralph Dum’s paper for the next session.

Some points to consider for GSS

To get the ball rolling…

• Systems approach to provide a framework for studying policy options, testing scenarios and decision making
• What needs to be considered and at what scale?
• Modelling techniques for analysis, predictions and forecasts.
• Incorporation of data – especially new data provided by user participation, data mining or ubiquitous computing
• Methods to offer new explanations, allow consultation and obtain feedback.
• Greater transparency and  relevance in periods of instability and uncertainty.

Workshops on Global Markets

Two small workshops on global markets and sustainability



Objects of governance have turned into parts of a globally interconnected system whose sustainability is at risk. This calls for a « global systems science», which could on the one hand support governance and, on the other hand, explicit how power and influence are transmitted in these interconnected systems. The aim of this workshop is to investigate how those two issues materialize in the economic realm: new questions and challenges economic policy addresses to science and new questions and challenges science faces when investigating economic policy and decision making in a globalized market environment.


Middle-term issues, whose time-span approximately coincides with the average five years electoral mandate, might possibly be dealt with by neo-classical social science: policy-makers maximize their probability of being reelected by putting in place policies whose efficiency (number of jobs created or quantity of pollution avoided) can be assessed “all other things being equal” using computable general equilibrium models.  Questions about sustainability are of a different kind, they are not asked “ceteris paribus” as they rely on endogenous dynamics and challenge the very existence of the system under investigation.


Sustainability mainly refers to the long-term balance between economic development and environmental conservation but also materializes as problems with very short time-horizons such as the 2011 earthquake in Japan, the failure of Lehman brothers, or the sovereign debt crisis in the Eurozone. Both for short and long-term issues, decision-makers turn to science with very specific as well with very general questions, with micro and macro problems. As far as environmental change is concerned, questions range from “How to reproduce locally an industrial symbiosis like this of Kalundborg ?”  to “How to move to a competitive low carbon economy in 2050 ?” Turning to the financial realm, one might wonder whether (or until when) the global financial system was sustainable or if it can today provide the 30 billions euros Greece needs by 2020 to upgrade its energy system[1] ?


It is doubtful that a single model might address those issues.  Even if one embeds all existing heterodoxies, very few of the policy-relevant questions of the time can be answered. Among the challenges to be faced in order to overcome this failure are the disaggregation and the extension to non-equilibrium phenomena of economic models and the exploitation of the wealth of data at hands thanks to the IT revolution.


The burgeoning mind of economists has managed to develop a wide taxonomy of models (e.g general equilibrium, CAPM, Bertrand or Cournot competition, principal-agent) under the, most often ad-hoc, constraint of only considering equilibrium situations. There is little doubt that this taxonomy can be extended if the economist toolbox is enriched by non-equilibrium models inspired by harder sciences.  How to foster this process ? For example how can econophysics and agent-based models become mainstream within the economic profession ? Which standards can be established to better communicate about models that aren’t always analytically tractable ?


New classes of models and new tools are also required to exploit the wealth of data at hands thanks to the IT revolution. Statistical institutes produce a limited and rather immutable class of indicators with a certain lag. Now, most if not all the data collected by statistical institutes during census operations are in fact readily available in some electronic form and circulating at extremely high speed on the internet. What are ethical and methodological issues posed by “big data” ?  Are current models up to the challenge of parsing if not explaining  the material ?



Finally, as a counterpart to the questions posed by policy, shouldn’t global systems science come with series of interrogations about power and governance in the globalized economy ? With markets’ liberalization and globalization, nation states seem to have lost part of the influence and the power they had on the economy. Where has this power gone ? Can citizens be offered better maps of the corporate governance network, the financial system or the international trade network ?  Aren’t such maps necessary to embed in the democratic process the redrawing of boundaries between the public and the private spheres, the public and the private sectors, communities and cities, membership and governance ?




Two series of questions



Global Markets I : Thursday, 8th November 2012, 15.15 -16.45


– What is unique about our system of globally interconnected markets that  materializes itself in this massive flow of data ? i.e “Big data” is the symptom of what ?

– How can this data be used to better understand the system which generates it?

– How can models parse and/or explain this  “big data” ?

– What is specific to models of the global economic system ?

– For example, does increasing connectivity lead to increasing fragility, what transpires in models as volatile or even chaotic behavior ?



Global Markets II: Friday, 9th November 2012, 10.30 -12.30:


How can  models using « big data » be more useful than existing ones for policy purposes ?

– What kind of models can be used for « emergency » economic and financial  management ?  By who and how ?

-What kind of models can be used for long-term issues such as strategic planning  or scenario generation ?  By who and how ?

– What kind of models do we need to better manage the transition to a green economy ?

-How can models improve public understanding of the dynamics of global systems ?

What would  be the impact of on governance ?

‘Global Systems Science’ Conference (Ralph Dum)

The conference will help ‘Global Systems Science’ to develop an agenda of action both in research (in ICT and beyond) and in policy (as a coordinating concept). GSS has as ambition to unite different strands of research and policies trying to cope with global challenges. The attached document is an attempt to do so, based on preceding consultations. The conference will ideally help to evolve and expand this preliminary document. (Ralph Dum)

Towards a ‘Global Systems Science’ : how this conference should contribute