Building a climate-neutral, green, fair and social Europe is one of the four current priorities for the strategic agenda for the European Union (EU). This is also present in one of the six priorities of "A European Green Deal", with the aim of making Europe carbon-neutral by 2050.
To reach this very ambitious aim, various actions should take place, like the decommissioning of fossil-based power plants in favor of renewable energy sources (RES) and the massive introduction of energy storage facilities, to capture electricity produced by RES and use it when these sources are not available.
In order for this process to remain resource-efficient and economically competitive, one has to determine, for each possible initiative, its expected future costs and benefits, to assess the global sustainability of this strategy. Mathematically, this typically translates into stochastic control and optimization problems, like e.g. optimal operation of power plants and/or storages and irreversible installation of RES, which are also useful to assess the effectiveness of various market mechanisms which are introduced into electricity market in order to facilitate this transition.
Moreover, many relevant problems central to energy transition are typically decentralized. For this reason, the related mathematical problems are naturally formulated in a multi-agent setting, possibly with a very large number of agents.
Besides the obvious involvement of member states, public regulatory agencies and leading energy producing industries, a key role of this energy transition will be played by the final consumers, who should be persuaded or pushed to change old habits, based on the existing fossil-fueled energy system, in favor of new habits: this could entail e.g. the adjustment of energy-consuming behaviors, the installation of individual RES-based assets (like photovoltaic panels on their roofs, or electricity storages) or the switch to electric cars.
Persuading final consumers and the general public to be more aware of climate change challenges and to have more energy-efficient behavior can be achieved via Bayesian persuasion, which is a set of modeling and mathematical tools which help to design optimal ways to release information to induce change in people habits and reduce negative externalities.
We aim to address these challenges using a large range of probabilistic techniques spanning stochastic control, filtering theory, games for finitely many players and games for a large population of players, such as mean field games, and cutting edge numerical methods based on neural networks. We plan to do so through three main working packages (WPs):
WP1: Bayesian persuasion for energy transition;
WP2: Optimal management of RES;
WP3: Mean field control problems and games for energy transition
