Dimitri Papageorgiou

Forged a highly-productive, cross-organizational team of researchers, analysts, and software engineers to identify in excess of $100M per year in business uplift. Defined a vision for the integrated team, connecting our strategic, tactical, and operational goals and deliverables. Guided team to develop a multiperiod blending analytical tool to help business pursue most profitable procurement strategy. Developed an optimization model and performed detailed analysis to determine potential biofuels pathways for supplying lower carbon intensity U.S. light-duty vehicle transportation fuels under different demand scenarios. Presented key findings to VP-level stakeholders and influenced strategic policy decisions.

Developed a first-of-its-kind assessment tool to evaluate technology trade-offs, emissions, and energy system pathways to enhance ExxonMobil's planning capabilities. The tool was recognized as a breakthrough technology for the corporation as it provides significant, immediate, and direct impact to VP-level managers. The energy system modeling capability "brings a new perspective to evaluate the research portfolio and strengthen guidance to the corporation about the role of technology in the energy transition." Recognized as an "Energy Pioneer" for providing the corporation with a decision support tool whose scale, complexity, flexibility, and robustness surpass those of existing tools.

Led team of PhDs, analysts, software engineers, and interns to improve ExxonMobil's market analysis and outlook capabilities by providing advanced optimization models and algorithms for robust analysis of market dynamics, including supply and demand elasticity. By adapting modern game theory and microeconomics, our team built innovative capability for ExxonMobil with the potential to dramatically improve our market behavior understanding, enhance our price forecasting, and ultimately inform multi-billion dollar resource acquisition, development, and marketing decisions.

Led collaborations with top universities including Carnegie Mellon University, Georgia Tech, National University of Singapore, MIT, University of Texas at Austin, and University of Toronto to understand the technical, operational, and policy requirements for increasing levels of intermittent renewables penetration. Research themes include, but were not limited to: Power system capacity expansion: Developed novel mathematical optimization formulations and algorithms to project long-term generation investment decisions, while accounting for important operational and chronological details. Explored various clustering techniques and optimization approaches for selecting representative days to approximate annual variability. Power grid operations: Investigated unit commitment and economic dispatch models to analyze hourly and subhourly electricity grid operations in the US (e.g., WECC and ERCOT) and India. Devised and implemented a clustering-based production cost simulation to analyze the impact of natural gas combined cycle generators with carbon capture and sequestration technology. Impact of EV charging on hourly operations and overall power grid carbon intensity.

Developed and patented sophisticated, but practical field-wide optimization models and algorithms to increase oil and gas production in large on- and off-shore assets. Co-developed information-theoretic sensor planning approach for large-scale production surveillance using deep reinforcement learning.

Led collaborations with two US universities to develop robust models for inventory management and vessel routing of crude tankers, LNG ships, and barges transporting refined products. Our results provided insight for future deliverables.

Advisor: George Nemhauser (Member of the National Academy of Engineering) Co-advisors: Joel Sokol (2011-12) and Martin Savelsbergh (2009-10) Participated on a project sponsored by ExxonMobil Corporate Strategic Research aimed at modeling and solving large-scale logistics problems.

Scientific staff member on multi-million dollar R&D projects guiding innovation in sensor resource management, decentralized decision-making, and multi-target tracking. As lead Operations Research analyst, oversaw the design, implementation, and analysis of tactical software used to improve solution speed and quality for track-to-track association problems. Developed optimization software for Raytheon's state-of-the-art multiple hypothesis tracker. Applied approximate dynamic programming methods to hierarchically schedule sensors to search for, track, and discriminate time-critical targets in a coordinated manner.

Developed tri-level network interdiction models to identify the most critical electric power grid components to defend from terrorist attacks and natural disasters. Designed and implemented specialized cutting-plane and decomposition algorithms to solve associated mixed-integer optimization models to optimality.

Constructed portfolio optimization models using mathematical programming to identify high payoff options for improving airport and aviation security after the 9/11 attacks. Results used in final publication and briefing to the U.S. Air Force. Created queuing simulations of airport security checkpoints to analyze efficiency of proposed passenger and baggage screening techniques.

Participant in the Langley Aerospace Research Student Scholars (LARSS) program. Developed a decision support prototype to provide project managers of NASA Langley’s Optimization branch with quantitative data for research prioritization and investment allocation. Incorporated stochastic models and performed Monte Carlo simulations with Excel and CrystalBall software.

Advisor: George Nemhauser (Member of the National Academy of Engineering) Co-advisors: Joel Sokol (2011-12) and Martin Savelsbergh (2009-10) Participated on a project sponsored by ExxonMobil Corporate Strategic Research aimed at modeling and solving large-scale logistics problems.

While working full-time at Raytheon, I took several graduate courses and participated in reading groups focused on Reinforcement Learning and Approximate Dynamic Programming.

Advisor: David Morton Formulated stochastic network interdiction models used for decision support in preventing and deterring the smuggling of nuclear material.

Graduated with Highest Distinction