Thermal storage facilities help to mitigate and to manage temporal fluctuations of heat supply and demand for heating and cooling systems of single buildings as well as for district heating systems. We focus on a heating system equipped with several heat production units using also renewable energies and an underground thermal storage. The thermal energy is stored by raising the temperature of the soil inside that storage. It is charged and discharged via heat exchanger pipes filled with a moving fluid.

Besides the numerous technical challenges and the computation of the spatio-temporal temperature dis- tribution in the storage also economic issues such as the cost-optimal control and management of such systems play a central role. The latter leads to challenging mathematical optimization problems. There we incorporate uncertainties about randomly fluctuating renewable heat production, environmental conditions driving the heat demand and supply.

The dynamics of the controlled state process is governed by a PDE, a random ODE, and SDEs modeling energy prices and the difference between supply and demand. Model reduction techniques are adopted to cope with the PDE describing the spatio-temporal temperature distribution in the geothermal storage. Finally, time- discretization leads to a Markov decision process for which we apply numerical methods to determine a cost-optimal control.

This is joint work with Paul Honore Takam (BTU Cottbus-Senftenberg) and Olivier Menoukeu Pamen (AIMS Ghana, University of Liverpool).

In this talk we present a novel deep learning methodology to detect financial asset bubbles by using observed call option prices. We start with an introduction to deep learning and asset price bubbles. We then examine the pitfalls of a naive approach for deep learning-based bubble detection and subsequently introduce our method. We provide theoretical foundations for the method in the context of local volatility models and show numerical results from experiments both on simulated and market data.

The talk is based on joint work with Francesca Biagini, Andrea Mazzon and Thilo Meyer-Brandis.

In this talk, we introduce two problems of contract theory, in continuous-time, with a multitude of agents. First, we will study a model of optimal contracting in a hierarchy, which generalises the one-period framework of Sung (2015). The hierarchy is modelled by a series of interlinked principal-agent problems, leading to a sequence of Stackelberg equilibria. More precisely, the principal (she) can contract with a manager (he), to incentivise him to act in her best interest, despite only observing the net benefits of the total hierarchy. The manager in turn subcontracts the agents below him. We will see through a simple example that, while the agents only control the drift of their outcome, the manager controls the volatility of the Agents’ continuation utility. Therefore, even this relatively simple introductory example justifies the use of recent results on optimal contracting for drift and volatility control, and therefore the theory on 2BSDEs. 

This will lead us to introduce the second problem, namely optimal contracting for demand-response management, which consists in extending the model by Aid, Possamai, and Touzi (2022) to a mean-field of consumers. More precisely, the principal (an electricity producer, or provider) contracts with a continuum of agents (the consumers), to incentivise them to decrease the mean and the volatility of their energy consumption during high peak demand. In addition, we introduce a common noise, impacting all consumption processes, to take into account the impact of weather conditions on the agents’ electricity consumption. This mean-field framework with common noise leads us to consider a more extensive class of contracts. In particular, we prove that the results of [1] can be improved by indexing the contracts on the consumption of one agent and aggregate consumption statistics from the distribution of the entire population of consumers. 

Talk based on Hubert (2020) and Elie, Hubert, Mastrolia, and Possamai (2021).

I will show that a zero-sum stopping game in continuous time with partial and/or asymmetric information admits a saddle point (and, consequently, a value) in randomised stopping times when stopping payoffs of players are general càdlàg adapted processes. We do not assume a Markovian nature of the game nor a particular structure of the information available to the players. I will discuss links with classical results by Baxter, Chacon (1977) and Meyer (1978) derived for optimal stopping problems. Based on a joint work with Tiziano De Angelis, Nikita Merkulov and Jacob Smith.

We present a stochastic Tikhonov theorem for two-scales systems of SDEs, which cover the case of McKean-Vlasov SDEs. Our approach extends and generalizes previous results on two-scales systems of SDEs without mean-field interaction. As an application we provide a novel method for approximating the solution of certain systems of FBSDEs, related to the Pontryagin maximum principle, which is new even for the case without mean-field interaction. This is a joint work with A. Cosso.

We consider the Bachelier model with linear price impact. Exponential utility indifference prices are studied for vanilla European options and we compute their non-trivial scaling limit for a vanishing price impact which is inversely proportional to the risk aversion. Moreover, we find explicitly a family of portfolios which are asymptotically optimal.

In this talk, we introduce several functional law of large numbers (FLLN) and functional central limit theorem (FCLT) for quasi-stationary Hawkes processes. Under some divergence conditions on triggered events, We prove that the normalized point processes can be approximated in distribution by a long-range dependent Gaussian process. Differently, both FLLN and FCLT fail when triggered events satisfy aggregation conditions. In this case, we prove that the rescale Hawkes process converges weakly to the integral of a critical branching diffusion with immigration. Also, the convergence rate is deduced in terms of Fourier-Laplace distance bound and Wasserstein distance bound. This talk is based on a joint work with Ulrich Horst.

This work is mainly concerned with the so-called limit theory for mean-field games. Adopting the weak formulation paradigm put forward by Carmona and Lacker, we consider a fully non-Markovian setting allowing for drift control, and interactions through the joint distribution of players’ states and controls. We provide first a new characterisation of mean-field equilibria as arising from solutions to a novel kind of McKean–Vlasov back- ward stochastic differential equations, for which we provide a well-posedness theory. We incidentally obtain there unusual existence and uniqueness results for mean-field equilibria, which do not require short-time horizon, separability assumptions on the coefficients, nor Lasry and Lions’s monotonicity conditions, but rather smallness conditions on the terminal reward. We then take advantage of this characterisation to provide non-asymptotic rates of convergence for the value functions and the Nash-equilibria of the N-player version to their mean-field counterparts, for general open-loop equilibria. This relies on new back- ward propagation of chaos results, which are of independent interest. This is a joint work with Ludovic Tangpi. 

The theory and applications of mean-field games/control have stimulated a growing interest and generated important literature over the last decade since the seminal papers by Lasry/Lions and Caines, Huang, Malhamé. This talk will address some learning methods for numerically solving  continuous time mean-field control problems, also called McKean-Vlasov control (MKV). In a first part, we consider a model-based setting, and present  numerical approximation methods for the Master Bellman equation that characterises the solution to MKV, based on the one hand on particles approximation of the Master equation, and on the other hand on cylindrical neural networks approximation of functions defined on the Wasserstein space. The second part of the lecture is devoted to  a model-free setting, a.k.a. reinforcement learning. We develop a policy gradient approach under entropy regularisation based on a suitable representation of  the gradient value function with respect to parametrised randomised policies. This study leads to actor-critic algorithms for learning simultaneously and alternately value function and optimal policies. Numerical examples in a linear-quadratic mean-field setting illustrate our results.