I'd agree going global is the right direction to take, it's the only time you really reach a truly closed system. I'd also agree that starting from a resources and a physical limits constraint is right (we're not likely to change these).

However I'd contend that as we hit peak/demand<>supply crossover the system enters a new phase and becomes highly complex, non-linear, and possibly out of control entirely. Taking nice proportional behaviours as a basis seems troubling.

The oil cost example is an interesting one. We know that price will rise to push out demand. However we also know that the impact of price rises on demand is connected to what the price history is, what else money is being spent on, which countries have the least leeway. Throw in some exportland, some 'non renegotiable' wars and some planned economic warfare and its difficult to know of oil will be $70 or $700 in only one year's time. I can make an argument for either. That in itself casts all other relationships in the model into question (can't scale solar PV if the economy collapses).

I'd be happier to see such a model married up to a 'wargame' with some really sneaky players (which is something I've been thinking about). That might delineate the parameters of the gamespace.

You mean combining it with a Department of Defense wargame simulation, like this?

Well, that's nice marketing spin for an agent based sim, but I was thinking more of sneaky humans working out the tricks of the new game scenario - akin to this:
http://www.guardian.co.uk/g2/story/0,3604,786992,00.html

We all know that one of the expected behaviours of a post peak world is for producers to artificially constrict supply even further, to push up prices and maintain reserves even longer. What other behaviours are likely? Are alternatives strangled at birth to keep the lights on today? What's credible with realistic decision types?

After all, the Easter islanders could have turned the last trees into boats and escaped, rather than burn them for firewood and to create statutes.

I have also given thought to an agent based modeling approach to understand the behavioral response (or lack thereof) to peak oil.

Think for a minute about the major players (agents) in such a model--and what drives them. There are oil companies, who seek to maximize shareholder value. Higher prices generate short-term profits, but long-term value depends on the value of their reserves.

Then there are oil producing (net exporting) countries, who depend on oil revenues to maintain their economies. Their access to oil also gives them a bargaining chip in the world, but also makes them vulnerable to invasion by large, oil consuming countries (i.e. the next agent described).

Major oil consuming (net importing) countries have their own set of agendas. Their economies are / were built by cheap oil. Without cheap oil, their massive economies come to a screetching halt. Leaders in these countries may (or may not) understand the gravity of the energy situation, however, are personally motivated to stay in power. The need for votes may contrain what they are able or willing to do.

Then there are investors, who seek to maximize short and long term financial returns. However, their decisions are only as good as the information they are provided. In the case of oil, much of this information is provided by estimates and forecasts by major oil companies, oil exporters and oil consuming nations.

For each of the described agents (and there are more), several key question stand out. 1. Are each of these agents aware of the peak oil threat? 2. If so, what is their likely response based on what motivates them? 3. How do the combined responses work together to explain our current situation? 4. What are the likely scenarios going forward?

Simple rules, but complex outcomes to be sure!

Debbie

Edit - I left out one of the most important agents--John or Jane Q Public.