6 Sep

When Controls meet Economics and Social Sciences

VeranstaltungsortLakeside B04, Eingang b, 1. Stock, Raum B04.1.114Veranstalter Institut für Vernetzte und Eingebettete SystemeLakeside Labs GmbHBeschreibungThe massive penetration of smart technology has changed energy systems and many other engineering applications such as transportation, logistics, and security. In all these contexts, one often observes a large number of machines and humans interacting with each other. Thus there is a need to redefine the boundaries of some engineering disciplines to accommodate both physical and socio-economic components. Control Engineering, the discipline which studies the ways in which one can control the evolution of a system is no exception. A core concept in control is ‘feedback’ whereby a machine (the controller) gets measurements from sensors about the state of the system one wishes to control and sets the inputs to the system. In the presence of numerous machines and humans, one observes the humans turning into strategic players who learn the environment and make decisions knowing that the environment (which is constituted by the other players) is at the same time learning about them and will react accordingly. This takes the concept of feedback to the next level and opens the floor to a number of game theoretic aspects (learning, incentives, pricing).In this tutorial I will present new perspectives and challenges arising when dealing with co-existing physical and socio-economic components. I will introduce dynamic games with a large number of players, (also known as mean-field games) and discuss recent trends. I will also discuss the role of strategic thinking and learning in competitive scenarios. The tutorial will conclude with a look at bio-inspired collective decision making problems and related evolutionary game models.Vortragende(r)Prof. Dario Bauso, Jan C. Willems Center for Systems and Control, ENTEG, Faculty of Science and Engineering, University of Groningen (The Netherlands), and Dipartimento di Ingegneria, University of Palermo (Italy)KontaktMelanie Schranz (schranz@lakeside-labs.com)

12 Sep
26 Sep

Backward Stochastic Differential Equations beyond Lipschitz Data

VeranstaltungsortN.0.07Veranstalter Institut für StatistikBeschreibungThe talk starts with a basic introduction to backward stochastic differential equations (BSDEs), then presents various applications of these equations in PDE theory, control theory, stochastics on manifolds and mathematical finance. Directly following will be an overview of the BSDEs' mathematical treatment with special focus on the very relevant case where the equations' data is irregular: data that satisfy a Lipschitz assumption lead to a well-behaved theory and many results are available. In many applications however, instead of Lipschitz conditions, only a quadratic growth or a mere local Lipschitz condition is available, rising questions that lack a big unifying theory so farVortragende(r)Dipl.-Ing. Alexander Steinicke, PhD (Montanuniversität Leoben)KontaktSimone Gahleitner (simone.gahleitner@aau.at)

7 Okt