Time inconsistent performance criteria are natural in optimization problems in themes such as large scale cooperative multi-agent systems including the principle-multi-agent (leader-follower) problem and crowd dynamics, diffusion of energy or fluids in non-homogeneous media, risk management of uncertain environment such as risks in river and coastal hydrology due to climate changes etc. I will comment on some challenges faced by modern optimization techniques when the performance criteria are time-inconsistent. I will show for example that time-inconsistency makes the Dynamic Programming Principle fail to hold even for the very simplest optimization problem and proposes a way to overcome this difficulty.
Biography: Boualem Djehiche is a Professor of mathematical statistics at the Division of Mathematical Statistics of the Department of mathematics, KTH, Sweden. His research interests are in the area of Stochastic Analysis and include the Theory of Large Deviations, Superprocesses and Interacting Particle Systems, with applications in Euclidean Quantum Mechanics, Insurance Mathematics, Mathematical finance and Mathematical Epidemiology.