<DrumBeat: February 16, 2008The Oil DrumDrumBeat: February 17, 2008>
<European Energy Conference, Aberdeen 22nd February 2008The Oil Drum: EuropeOilwatch Monthly - February 2008>

244 comments on Peak Oil and "The Limits to Growth": two parallel stories

Comments can no longer be added to this story.

galacticsurfer on February 12, 2008 - 3:55pm Permalink | Subthread | Comments top

System dynamics sounds great. Working upwards and outwards or staying in a small system, being modular, all depending on what you intend to model, the world or a limited system. If we at TOD want to get away from just resource modeling with Hubbert Linearization and talk about the big picture wtih various planning scenarios this is the set of tools we will have to use. For a layman or anyone with a good science or engineering background and an inteest it seems this would be what we need to learn in our spare tim to figur out how to see what is possible and what not. I would presume however that such a model, depending on what one has inmind needs a computer programme, custom made and lots of data gathered over many months so that it would be impractical for our off th cuff threads made over a weekend. Is this right or can one just work up something quick and dirty with standard tools on the web or elsewhere?

0
Comments can no longer be added to this story.
Ugo Bardi on February 13, 2008 - 4:39am Permalink | Subthread | Parent | Comments top

I've found that learning system dynamics is a little like learning to play chess. You first learn how to move pieces, then you can play the game. But becoming a master takes years of practice. I am still a novice in system dynamics; I am learning. I can make models using Vensim and found that it is a good way for focusing one's mind on the way the system you are studying works. On the other hand, coming from Hubbert style modeling, I found surprising that Vensim (and not even Stella, as far as I know) has no routines for data fitting. It is a different world, a different philosophy. S.D. models, it seems, are not supposed to be predictive tools. Nevertheless, I am trying to do exactly that; use S.D. models to fit oil production curves. For that, I had to enlist the help of a coworker who is a better programmer than I am and who built specific routines using Simulink. We are getting interesting results that we plan to publish, but we are doing that in our spare time, as usual. It is a law of academia that the most interesting research project is the one that doesn't get financial support. So, it moves on, but sloooowly......

0
Comments can no longer be added to this story.
garyp on February 13, 2008 - 4:47am Permalink | Subthread | Parent | Comments top

Systems dynamics has its uses (IIRC that's what the World Energy Modeling Project are using). However it tends to break down in two main areas:

  1. Sharp discontinuous change where things switch from one model to another
  2. Human beings

So when you are talking about resources or areas where humans just react blindly it can deliver some insight. However when things start going wrong it can badly fail to represent strategic or tactical action from C&C type economies.

As a simple for instance. It would be perfectly possible to imagine someone letting off a nuke in the heart of the Saudi oil region. The instantaneous effect on oil supply would be obvious, as would the environmental impacts. However it would also bring other worldview models to the front. The world after such an event wouldn't run on the same rules as the world before. There is no way system dynamics can model such changes since the very makeup of the model depends on your understanding of how the world works now.

If you want to play, there are SD tools around. However I'd tend to focus on complex adaptive systems approaches if I were you.

0
Comments can no longer be added to this story.
galacticsurfer on February 13, 2008 - 5:30am Permalink | Subthread | Parent | Comments top

I googled complex adaptive systems and got a university link with some Mac software:

http://cognitrn.psych.indiana.edu/rgoldsto/complex/

and a quote:

Students should gain both an intuitive appreciation for the behavior of complex adaptive systems as well as an understanding of their formal underpinnings.

An active exploration of complex systems is the only true way to come away with a practical understanding of them. In many cases, it is quite surprising to see the highly organized, orderly behavior that emerges from systems composed out "agents" behaving in accord with very simple rules. Many people who have worked with these computer simulations for an extended period of time have been struck by the qualitative increase in their knowledge, relative to simply seeing the equations underlying the systems or a verbal description.

Simulations should show how agents organize themselves dynamically, over space and time.

Students should be able to actively change the behavior of the systems by "tweaking" parameters. All of the simulations that I have written contain a substantial amount of flexibility in them. Students can feel in control of the simulation, and can discover new phenomena and interrelations by systematically changing parameters. Students can conduct experiments, and instantly see the results of the experiments. In addition, the interactivity makes the simulations more appealing to use, hopefully increasing the amount of time students will spend with them.

http://www.innovation.cc/volumes-issues/rogers-adaptivesystem7final.pdf

The Complex Adaptive Systems (CAS) Model was born of the scientific study of complexity. According to James Gleick, the inspiration for complexity science can be traced to John von Neumann’s dynamic weather system models of the 1950s at the Institute for Advanced Study in Princeton, New Jersey, an effort that, in turn, goes back to the work of the eighteenth-century philosopher-mathematician Laplace (Gleick, 1987, p. 14). The diffusion of innovationsmodel, credited to Everett Rogers, delineates the process by which an innovation spreads via certain communication channels among members of a social system (Rogers, 2003). Diffusion phenomena bear a resemblance to complex adaptive systems.

“In linear systems the relationship between cause and effect is smooth and proportionate. Linear systems respond to big changes in a big and proportionate manner and linear systems respond to small changes in an equally small and proportionate way” (Kiel, 1995). Most real life situations, on the other hand, are complex. Small changes in initial conditions, and later interventions of whatever size, can result in disproportionately large effects.

and from WIKI of course:

http://en.wikipedia.org/wiki/Complex_adaptive_system

A CAS is a complex, self-similar collection of interacting adaptive agents. The study of CAS focuses on complex, emergent and macroscopic properties of the system. Various definitions have been offered by different researchers:

John H. Holland
A Complex Adaptive System (CAS) is a dynamic network of many agents (which may represent cells, species, individuals, firms, nations) acting in parallel, constantly acting and reacting to what the other agents are doing. The control of a CAS tends to be highly dispersed and decentralized. If there is to be any coherent behavior in the system, it has to arise from competition and cooperation among the agents themselves. The overall behavior of the system is the result of a huge number of decisions made every moment by many individual agents.[1]

Kevin Dooley
A CAS behaves/evolves according to three key principles: order is emergent as opposed to predetermined (c.f. Neural Networks), the system's history is irreversible, and the system's future is often unpredictable. The basic building blocks of the CAS are agents. Agents scan their environment and develop schema representing interpretive and action rules. These schema are subject to change and evolution.[2]

Other definitions
Macroscopic collections of simple (and typically nonlinearly) interacting units that are endowed with the ability to evolve and adapt to a changing environment.[3]

0
Comments can no longer be added to this story.
garyp on February 13, 2008 - 6:16am Permalink | Subthread | Parent | Comments top

Sorry, I did tend to throw away the end comment.

The benefit from my perspective is that you can create individual agents which behave as actors in your simulation, complete with known behaviours, possible behaviours, and memory. You then combine these and look at what comes out of the complex whole, together with a degree of Monte Carlo simulation of parameters, random events. From the population of results you gain understanding of how the simulation will tend to react across a likely range of circumstances (eg where are the attractors) and if your simulation is anything like reality, a deeper understanding of how real systems will react. Agents don't need to be people/nations, they can be identified groupings, resources, anything.

I think you can see from the links you pulled out why I think they are a better match for modelling the characteristics we are interested in.

0
Comments can no longer be added to this story.
Lucifer on February 13, 2008 - 6:35am Permalink | Subthread | Parent | Comments top

intuitive appreciation !

Studied all this stuff (and much more) through those, you know, "Quantum Chromodynamics & The Charmed Quark for Dummies" style books.(populist commentries may be the precise term ?)

These matters you good folks speak of are definitely valid as methodologies and I basically make all my decisions in life based upon an intuitive mathematical engine that compiles many varieties of maths, science and pseudo-science into a functional world view. I'm doing very well thanks. A little bird tells me, WE are not doing so well at all.

Prediction - things are going to "blow up in our face" within weeks.
(The Finance system is an odds on favourite, so it probably won't be that LOL)

Smile, it's a good feeling.

0
Comments can no longer be added to this story.