The least-worst regret approach to grid investment under uncertain EV and heat pump adoption
What you’ll learn
- Apply the minimax regret framework to investment decisions under deep uncertainty.
- Construct a regret matrix to compare candidate plans against plausible future scenarios.
- Identify the plan with the smallest worst-case regret and justify it to stakeholders.
- Distinguish between ordinary uncertainty and deep uncertainty, and know when each planning approach applies.
Course Content
- Introduction –> 1 lecture • 4min.
- Description of the case –> 3 lectures • 13min.
- Minimax Regret –> 4 lectures • 39min.
- BONUS –> 1 lecture • 1min.
Requirements
WHO I AM: I am Dr. Spyros Giannelos. I hold a PhD in Energy from Imperial College London, and I am the founder of The Energy Data Scientist Academy. I teach practical, real-world data science specifically for the energy sector, and my research has received over 1300 citations globally.
REGULAR ENHANCEMENTS: This course is reviewed periodically with updates to reflect the modern energy market.
STUDENT BONUS: Note: Students who enroll in this course will receive access to 1 additional online course (industry case study).
What You’ll Learn:
- How to make investment decisions when you cannot assign probabilities to future scenarios
- The minimax regret framework from decision theory, applied to real infrastructure planning
- How to build a regret matrix that compares every candidate plan against every possible scenario
- How to identify the plan with the smallest worst-case regret, step by step
- The strengths and limitations of minimax regret compared to expected-value approaches
- How to present results to non-technical stakeholders including regulators and executives
Perfect For:
- Analysts in regulated industries facing long-horizon investment decisions
- Infrastructure planners (energy, water, transport) working under structural uncertainty
- Consultants advising utilities, governments, or European institutions on investment strategy
- Students in energy economics, operations research, or decision science
- Professionals who need a defensible framework for decisions where probabilities are contested
Why This Matters:
Most investment planning tools assume you can assign probabilities to future outcomes. In practice, for decisions involving electric vehicle adoption, heat pump rollout, rooftop solar, or geopolitical shifts, probabilities cannot be defended. Minimax regret offers a principled answer: it asks which plan you will be least sorry about, no matter which future arrives.
This course walks through the framework using a real distribution network planning example with three candidate plans and three scenarios. You will see exactly how the regret matrix is constructed, how the minimax solution is identified, and why it often produces the middle-ground plan that is robust across all plausible futures.
The course is focused on applied reasoning rather than abstract theory. It uses numerical examples throughout and explains every concept plainly, without assuming a background in decision theory or optimisation.