Georgios Pantazis

Research

Decision making and control under uncertainty from limited data

My research studies how to make reliable decisions when uncertainty cannot be modeled accurately in advance and must instead be inferred from data. A central question in my work is: what can we guarantee when the uncertainty model itself is only partially known?

I develop methods for optimization, games, and control that use finite data to quantify uncertainty and establish guarantees on properties such as feasibility, stability, and robustness. I am particularly interested in systems with multiple interacting decision makers and, more recently, in networked systems operating with imperfect or corrupted information.

Research Areas

Data-driven games and multi-agent decision making

Many engineering systems involve interacting decision makers operating under uncertainty. I develop data-driven and distributionally robust methods for multi-agent optimization and equilibrium problems, with particular emphasis on what finite data can guarantee about Nash and generalized Nash equilibria when agents face heterogeneous uncertainty.

Selected publications

A priori data-driven robustness guarantees on strategic deviations from generalised Nash equilibria, Automatica, 2024.
Probabilistically robust stabilizing allocations in uncertain coalitional games, IEEE Control Systems Letters, 2022.
Wasserstein Distributionally Robust Nash Equilibrium Seeking with Heterogeneous Data: A Lagrangian Approach, IFAC World Congress, 2026

Control under uncertain and corrupted data

Control decisions increasingly rely on data describing disturbances, operating conditions, and the surrounding environment. I study how uncertainty can be quantified directly from limited or imperfect data and incorporated into control and estimation methods, with the aim of providing rigorous guarantees without assuming that the underlying uncertainty distribution is known exactly.

Current directions include data-driven reachable set estimation, robust and resilient control under compromised measurements, and control of dynamical systems under uncertain operating conditions.

Selected publications

Data-driven Reachable Set Estimation with Tunable Adversarial and Wasserstein Distributional Guarantees, IEEE CDC 2026.
Certificate-based Synthesis of Coordinated Droop Control for Heterogeneous Radial Distribution Networks, under review.

Energy systems and markets

Energy systems bring together many of the questions and challenges that motivate my research involving heterogeneous decision makers, shared physical infrastructure and uncertain operating conditions. I use optimization, game theory, and control to study how flexible resources can be coordinated reliably, with applications including electrical distribution networks, vehicle-to-grid participation and virtual energy storage systems.

Selected publications

A User-centric Game for Balancing V2G Benefits with Battery Degradation of Electric Vehicles, IEEE Transactions on Transportation Electrification, 2025.
A Game-theoretic Model for Flexible Heterogeneous Distributionally Robust Virtual Energy Storage Services, under review.
Distributionally Robust Virtual Energy Storage Services based on Parked and Idling Electric Vehicles, ICRA, 2026.