Miguel Lejeune
George Washington University, DC·Engineering
Power grids, roads and other infrastructure face growing threats from natural hazards such as wildfires. Protecting them costs a great deal and can take years.
- Fields
Engineering
Mathematics & Statistics
Computer Science
- Topics
operations research
stochastic optimization
stochastic programming
- Grants
- 2 known grants, through Jul 2028
What they work on
- Advance real options analysis for infrastructure protection investment under multiple sources of uncertainty, building on optimal stopping theory and the Black-Scholes model from finance.
- Model risk tradeoffs and decision-makers' risk behavior in hazard-protection planning, using chance-constrained stochastic integer programming via p-efficiency.
- Capture network-wide impacts and interacting investment options across infrastructure systems, using approximate dynamic programming, partially observable Markov decision processes, deep reinforcement learning and multi-agent extensions.
- Develop 'learn-to' algorithms that embed machine learning inside optimization for speed, and explainable AI that turns model results into actionable protocols for decision-makers.
- Build risk-averse stochastic programming models with adjustable risk aversion in which wildfire-prevention decisions change outcome probabilities (decision-dependent uncertainty), to improve power grid resilience.
- Design scalable reformulation and solution algorithms for decision-dependent, risk-averse grid optimization problems, so operators can respond to wildfire threats with minimal investment.
More about the research
Doing nothing, however, risks economic losses, physical damage, deaths and displaced communities. Decision-makers must choose when and how much to invest while the future is highly uncertain. Both projects build mathematical optimization methods for making these choices under uncertainty and risk.
Why it matters
Communities depend on infrastructure that natural hazards can knock out. Utilities and local decision-makers must spend large sums on protection without knowing exactly what is coming. These methods could help them decide when to invest in protecting civil infrastructure, weighing risk across whole networks. They could also help power grid operators prepare for wildfires at lower cost, reducing outage risk, economic loss and damage.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Chi Jin
Princeton University, NJ·Computer Science·Early-career
Identify the general structures of function classes, such as those represented by neural networks, under which model-free reinforcement learning is tractable, and design provably efficient algorithms that also handle model misspecification.
Emiliano Dall'Anese
Boston University, MA·Engineering·Early-career
Design real-time controllers that approximate optimization-based safety filters using only local or limited measurements, keeping networked systems with partially known dynamics within operating limits.
Iakovos Toumazis
University of Texas MD Anderson Cancer Center, TX·Medicine & Clinical Research
Expand the ENGAGE lung cancer screening framework beyond age, sex and smoking history to include non-smoking risk factors such as race, family history and history of pulmonary disease.
The school
George Washington University: 13 labs on PI Pages, $382M a year on research, 225 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.