**Past events:**

* ** *

**Optimal insider control of stochastic partial differential equations,with applications to optimal harvesting and optimal insider portfolio under noisy observations*** (ORCOS Seminar)*** **

by Prof. Bernt Øksendal (Department of Mathematics, University of Oslo)* *

*Date: * 9^{th} May 2017 - *Time: *16:30 - 18:00

*Location: *seminar room (DB gelb 04)

TU Vienna

wing B (yellow area), 4th floor

Wiedner Hauptstraße 8

*Abstract:*

We study the problem of optimal control with inside information of an SPDE (a stochastic evolution equation) driven by a Brownian motion and a Poisson random measure. Our optimal control problem is new in two ways:

(i) The controller has access to inside information, i.e. access to information about a future state of the system,

(ii) The integro-differential operator of the SPDE might depend on the control.

In the first part of the paper, we formulate a sufficient and a necessary maximum principle for this type of control problem, in the following two cases:

(a) The control is allowed to depend both on time t and on the space variable x.

(b) The control is not allowed to depend on x.

In the second part of the paper, we apply the results above to the problem of optimal control of an SDE system when the inside controller has only noisy observations of the state of the system. Using results from nonlinear filtering, we transform this noisy observation SDE inside control problem into a full observation SPDE insider control problem. The results are illustrated by explicit examples.

The presentation is based on joint works with Olfa Draouil, University of Tunis El Manar, Tunisia.

* ** *

**Optimized train speed profiles **

(PLIS/ PROLOG Colloquium)** **

by David Pisinger (DTU)* *

*Date: * 05.04.2017 - *Time: 17:00*

*Location: * SE 15

Faculty of Business, Economics and Statistics

(Oskar-Morgenstern-Platz 1, 1090 Vienna)

*Abstract:*

We present a novel solution method to generate energy-efficient train speed profiles. As opposed to previous analytical methods, we propose a graph representation of the problem, making it possible to generate pareto-optimal train speed profiles by use of Dynamic Programming (DP). Instead of using uniform discretization of space, time or speed, we rely on an event-based decomposition that drastically reduces the search space. Moreover, we are able to handle speed limitations, passage point time windows, as well as various measures of robustness (buffer time, number of speed changes). Such additional constraints were difficult to handle by previous approaches. Based on an extensive number of real-life problem instances our benchmark shows that the proposed solution method is able to reduce the energy consumption by 3.3% on average compared to existing solutions. The computational times are very low, making it possible to handle unexpected changes in speed restrictions or timings by recalculating the schedule. This is a great advantage compared to static offline lookup tables.

(co-authors: Jørgen Thorlund Haahr)

* ** *

**Differential Games with incomplete asymmetric information on the initial condition **

(ORCOS Seminar)** **

by Prof. Marc Quincampoix (Universite de Bretagne Occidentale)* *

*Date: * 04.04.2017 - *Time: 16:30-18:00*

*Location: * seminar room (DB geld 04)

(Wiedner Hauptstrasse 8, wing B (yellow area), 4th floor)

TU Wien

*Abstract:*

We investigate a two-players zero-sum differential game with an incomplete information. The first player has a complete information on the initial state of the game while the second player as only an information of probabilistic nature : he knows a probability measure on the initial state. The existence of a value for such game was obtained only when the probability measure has a finite support. Such differential games with finite type incomplete information can be viewed as a generalization of the famous Aumann-Maschler theory for repeated games.

The main goal and novelty of the present work consists in obtaining and investigating a Hamilton Jacobi Isaacs Equation satisfied by the upper and the lower values of the game. Since we obtain an uniqueness for such Hamilton Jacobi equation, as a byproduct, this gives the existence of a value of the differential game. Since the Hamilton Jacobi equation is naturally stated in the space of probability measures, we use the Wasserstein distance and some tools of optimal transport theory.

* ** *

**Mathematical optimization approaches for facility layout problems ***(ISOR Colloquium)*** **

by Miguel F. Anjos (PolyU Montreal)* *

*Date: * 20^{th} March 2017 - *Time: *16:45 - 17:45

*Location: * Hörsaal 7

Faculty of Business, Economics and Statistics

(Oskar-Morgenstern-Platz 1, 1090 Vienna)

*Abstract:*

Facility layout problems are an important class of operations research problems that has been studied for several decades. Most variants of facility layout are NP-hard, therefore global optimal solutions are difficult or impossible to compute in reasonable time. Mathematical optimization approaches that guarantee global optimality of solutions or tight bounds on the global optimal value have nevertheless been successfully applied to several variants of facility layout. In this talk, we review three classes of layout problems, namely row layout, unequal-areas layout, and multifloor layout, and summarize the recent developments in applying mathematical optimization to these classes. We also present some of our latest research results. We also briefly discuss directions that remain open for future research.

**Kick-Off Event:**

*Date: * 13^{th} March 2017 - *Time: *15:00

*Location: * Hörsaal 5

Faculty of Business, Economics and Statistics

(Oskar-Morgenstern-Platz 1, 1090 Vienna)

*Guest:* Christian Blum, PhD (Artificial Intelligence Research Institute & Spanish National Research Council )

*Topic: *CMSA: Taking profit from ILP solvers in the context of large problem instances

*Abstract:*

The combination of metaheuristics with complete techniques, such as integer linear programming (ILP) solvers, is one of the current lines of research in combinatorial optimization. The main aim behind such approaches is to exploit the complementary character of different optimization strategies in order to obtain robust algorithms that generate high-quality solutions in reasonable computation times. Given a combinatorial optimization problem, general purpose ILP solvers such as CPLEX are often highly efficient for solving instances up to a certain size which is problem-dependent. This is because they are built upon many years of research and they make use of efficient implementations of cutting edge ILP technologies. One of the motivations for the combination of metaheuristics with ILP solvers such as CPLEX is the aim of taking profit from the application of ILP solvers even when the considered problem instances are too large for applying these solvers directly. In this keynote talk we will present our most recent algorithmic development that resulted from the motivation outlined above: construct, merge, solve & adapt (CMSA). Moreover, we present our initial investigation on the relation of CMSA with large neighborhood search (LNS), which is one of the standard ways of combining local search with ILP solvers.

University of Vienna

Oskar-Morgenstern-Platz 1

A-1090 Vienna

T: +43-1-4277-38092

F: 43-1-4277-838094