Introducing uncertain learning in an energy system model: a pilot study using GENIE

An economic perspective on experience curves and dynamic economies in renewable energy technologies

“…However, the experience curve results in this paper indicate that although single experience indices may be highly statistically significant, whether estimated for individual technologies and countries or combined in panels, most of these estimates become insignificant when a time trend is included as an explanatory variable. These results support the introduction of imperfect foresight and stochastic uncertainty of learning rates in energy system models (Mazzola and McCardle, 1996;Matsson, 2002;Grubb et al, 2002).…”

Short term decisions for long term problems – The effect of foresight on model based energy systems analysis

“…These trends remain, no matter whether a moving window or a two step approach is used, or whether the perfect foresight case has flexible growth constraints. 9 These results again emphasize, how in the limited foresight cases the decisions based on a short to mid-term view alone, can lead to a lock-in, or a lock-out, which may increase the costs in the long term. Previous approaches have also made this very same conclusion [14].…”

“…Cumulative global fossil and other energy investments in the short, mid and long term. 9 However, costs are naturally increased, if no flexible growth constraints are available. 10 This price reflects only the extraction and distribution costs of the fuel that are represented in the model and does not therefore take into account many of the drivers that affect the high oil prices of the past years.…”

“…In addition to models that limit foresight mostly by adding stochasticity and/or known branching points (e.g. [9][10][11][12]. See also [13].…”

The Impact of Uncertainty in Climate Targets and CO2 Storage Availability on Long-Term Emissions Abatement

“…Hereby the future can be described in terms of alternative "states-of-the-world", of which the branching points, likelihoods and characteristics are known entities ( [18], for ETSAP-TIAM; see also [22], for another approach). The objective function is altered for this purpose, so that the costs of alternative states, each holding a distinct set of result variables after the branching point, are weighted according to their probabilities.…”