On Thursday last week my PhD student Kristofer Jakobsson defended his licentiate thesis, half way to the PhD thesis. The summary is given below and the full thesis you can download from our webpage. (To Global Energy Systems for download). The opponent was Robert Hart, Department of Economics, Swedish University of Agriculture. His final comment was that the thesis was very good.
Energy is essential to the functioning of society, and oil is the single largest commercial energy source. Some analysts have concluded that the peak in oil production is soon about to happen on the global scale, while others disagree. Such incompatible views can persist because the issue of “peak oil” cuts through the established scientific disciplines. The question is: what characterizes the modeling approaches that are available today, and how can they be further developed to improve a trans-disciplinary understanding of oil depletion? The objective of this thesis is to present long-term scenarios of oil production (Paper I) using a resource-constrained model; and an agent-based model of the oil exploration process (Paper II). It is also an objective to assess the strengths, limitations, and future development potentials of resource-constrained modeling, analytical economic modeling, and agent-based modeling.
Resource-constrained models are only suitable when the time frame is measured in decades, but they can give a rough indication of which production scenarios are reasonable given the size of the resource. However, the models are comprehensible, transparent and the only feasible long-term forecasting tools at present. It is certainly possible to distinguish between reasonable scenarios, based on historically observed parameter values, and unreasonable scenarios with parameter values obtained through flawed analogy. The economic subfield of optimal depletion theory is founded on the notion of rational economic agents, and there is a causal relation between decisions made at the micro-level and the macro-result. In terms of future improvements, however, the analytical form considerably restricts the versatility of the approach.
Agent-based modeling makes it feasible to combine economically motivated agents with a physical environment. An example relating to oil exploration is given in Paper II, where it is shown that the exploratory activities of individual agents can yield a U-shaped exploration cost path. Agent-based modeling appears to have significant potential for future development, but it is still unclear whether it will be the most useful in policy evaluation or more generalized systems research.
If you are very interested in the subject you can listen to a recording of the presentation in 4 parts: