Sequential decision-making under uncertainty for long-term energy transition planning

Global warming concerns have led to emission regulations and various incentives for low-carbon technologies. Energy system models, which are used to examine how investments affect our ability to meet energy demand, are typically based on two assumptions: key parameters are assumed to be known deterministically and a multi-period energy transition plan is determined at one point in time. We argue that for a systematic generation and analysis of energy transition pathways, these assumptions should be relaxed and, accordingly, we propose methods to achieve that. First, we use stochastic programming (SP) to account for uncertainty in key parameters. Second, we pair SP with a sequential decision-making approach that represents how decisions can be updated as uncertainties unfold. Third, we use simulation-based methods to evaluate the quality of energy transitions. Importantly, we find that accounting for uncertainty, proactively and through feedback, yields pathways with diverse technology portfolios that are resilient to uncertainty.