From 14th to 16th June 2021
Dubrovnik, Croatia

Participants

Invited speakers

Peter Benner. Max Planck Institute for Dynamics of Complex Technical Systems, Germany.
Title: LQResNet: Using DNNs for Learning of Dynamical Systems
(abstract).

Marko Erceg. Department of Mathematics, University of Zagreb, Croatia.
Title: pending
(abstract).

Luka Grubišić. Department of Mathematics, University of Zagreb, Croatia.
Title: Rational function surrogate modeling for the optimal control of parabolic systems
(abstract).

Andrej Jokić.
Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb, Croatia.
Title: Market-based power systems: a control perspective
(abstract).

Jérôme Lohéac.
CNRS, CRAN, Université de Lorraine, France.
Title: Some contributions to output controllability
(abstract).

Darko Mitrovic.
Faculty of Mathematics, University of Vienna, Austria.
Title: Global Controllability for Quasilinear Non-negative Definite System of ODEs and SDEs
(abstract).

Ivan Veselić.
Fakultät für Mathematik, Technische Universität Dortmund, Germany.
Title: Null controllability for the semigroup of the harmonic oscillator
(abstract).

Enrique Zuazua.
Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) / Alexander von Humboldt Foundation, Germany
/Fundación Deusto and Universidad Autónoma de Madrid, Spain.
Title: Dynamics and control of a guiding model
(abstract).

Travel and accommodation

Covid19 information:
As the time of your arrival to Dubrovnik is approaching, I would like to share with you further instructions
for crossing the Croatian border, in regards to situation with COVID pandemic. As the situation is not
yet under the control, there are still some restrictions you have to be aware of. You would have to present a
negative PCR test (not older than 48h) to be able to enter Croatia. Detailed instructions you can find
here.

It is also strongly recommended that all foreign citizens announce their arrival information
via web form “Enter Croatia”:
https://entercroatia.mup.hr/

Travel

The Dubrovnik International Airport is well connected to many major cities of the world, either directly or
by the international airports in Zagreb and Split.

Regular bus service by Atlas buses runs from the airport to the central bus station in Gruž. Buses depart
after each scheduled flight arrival, and the ticket costs 35 KN (or 5 EUR). The drive takes about 30 minutes.
On the way buses stop at Pile (at the entrance into the Old City), from where you can take a local city bus
(line 6) to the Conference venue, University of Dubrovnik, Department of Electrical Engineering and Computing.
Address:
Ćira Carića 4, 20000 Dubrovnik.

Please see a map city with all the building market,
link.

Buses heading for the airport leave from the central bus station 90 minutes before scheduled Croatia
Airlines or Austrian Airlines flight departures, and 120 minutes before scheduled departures of other airlines.
They also make one stop near the Old City, next to the cable car station.

You can also take a taxi from the airport (cost €40 approximately).

Additional information about getting by plane:
Dubrovnik airport
Croatia airlines

Accommodation

All participants are expected to make their own accommodation arrangements.

We recommend affordable accommodation in a separate building of student dormitory which is located in the center of Dubrovnik,
on address Marka Marojice 2b. In the dormitory there would also be a gym,
music club, caffee bar etc. Nearby there is a few small shops, supermarkets
are situated 10-15 minutes by walking from the dormitory.
Website.

Prices:

price per person in double room per day, breakast incuded: 285 kn + Tourist tax 20 kn per person per day.
price per person in double room (single use) per day, breakast incuded: 473 kn + Tourist tax 20 kn per person per day.

For more information please contact

Below we list few hotels in the nearby area:

https://www.valamar.com/en/hotels-dubrovnik (a resort, with hotels ranging 3-5*)

http://www.grandhotel-park.hr/ (4*)

https://www.hotel-lapad.hr/hr/contact-us/ (4*)

https://www.hotel-ivka.com (3*)

There are some pages you can also check and see some apartments to rent:
Airbnb
Booking

Schedule and abstracts

Schedule

14th June

Time (CET)	Speaker	Talk titles
Session 1 – Chair: name
08:55-09:00	Opening
09:00-09:45	Speaker 1	title
09:45-10:30	Speaker 2	title
10:30-11:00	Coffee break
11:00-11:30	Speaker 3	title
11:30-12:00	Speaker 4	title
Session 2 – Chair: name
15:00-15:45	Speaker 5	title
15:45-16:30	Speaker 6	title
16:30-17:00	Coffee break
17:00-17:30	Speaker 7	title
17:30-18:00	Speaker 8	title

15th June

Time (CET)	Speaker	Talk titles
Session 1 – Chair: name
08:55-09:00	Opening
09:00-09:45	Speaker 1	title
09:45-10:30	Speaker 2	title
10:30-11:00	Coffee break
11:00-11:30	Speaker 3	title
11:30-12:00	Speaker 4	title
Session 2 – Chair: name
15:00-15:45	Speaker 5	title
15:45-16:30	Speaker 6	title
16:30-17:00	Coffee break
17:00-17:30	Speaker 7	title
17:30-18:00	Speaker 8	title

16th June

Time (CET)	Speaker	Talk titles
Session 1 – Chair: name
08:55-09:00	Opening
09:00-09:45	Speaker 1	title
09:45-10:30	Speaker 2	title
10:30-11:00	Coffee break
11:00-11:30	Speaker 3	title
11:30-12:00	Speaker 4	title
Session 2 – Discussion

 

Abstracts

Peter Benner,
LQResNet: Using DNNs for Learning of Dynamical Systems

Mathematical modeling is an essential step, for example, to analyze the transient behavior of a
dynamical process and to perform engineering studies such as optimization and control. With
the help of first-principles and expert knowledge, a dynamic model can be built. However,
for complex dynamic processes, appearing, e.g., in biology, chemical plants, neuroscience,
financial markets, this often remains an onerous task. Hence, data-driven modeling of
the dynamics process becomes an attractive choice and is supported by the rapid advancement
in sensor and measurement technology.
Data-driven methods such as Dynamic Model Decomposition (DMD) and Operator Inference (OpInf),
learn models of dynamical systems from data. In this talk, we suggest combining OpInf with
deep neural network (DNN) approaches to infer the unknown nonlinear dynamics of the system.
The approach uses recent advancements in deep learning and prior knowledge of the process if
possible. We demonstrate that the proposed methodology accomplishes the desired tasks for
dynamics processes encountered in neural dynamics and the glycolytic oscillator.

This talk is based on joint work with Pawan Goyal.

Marko Erceg,
pending

Luka Grubišić,
Rational function surrogate modeling for the optimal control of parabolic systems

We consider an optimal control problem for a general linear parabolic equation governed by a self-adjoint
operator on an abstract Hilbert space. The task consists in identifying a control
(entering the system through the initial condition) that minimises the distance of the
trajectory of the system from a given constraint while steering the final state at
time T > 0 close to the given target. First, we obtain the closed form solution for several
types of trajectory constraints. Then second we present a numerical scheme based on the award
wining rkfit algorithm for approximating the trajectory
of the system using spectral calculus and a rational function surrogate.

This talk is based on joint work with Martin Lazar, Ivica Nakic and Martin Tautenhahn.

Andrej Jokić,
Market-based power systems: a control perspective

Over the past decades, electrical power systems have been going through some major
restructuring processes. Central to these changes are a policy shift towards competitive
market mechanisms for their operation and large-scale integration of intermittent, renewable
power sources (wind and sun). As a consequence, these systems are characterized with decentralized
decision making, large uncertainties, interconnected dynamics and large scale. The main focus
of this talk is on challenges in managing complexity of such a system in the operational and
control design, with an emphasis on scalability of analysis/synthesis, robustness and reliability.
Several results will be presented and discussed, including distributed, real-time price-based
control of power systems, provision of power balancing ancillary services and distributed
algorithms for robust congestion management.

Jérôme Lohéac , Some contributions to output controllability

In this talk we will only deal with linear time invariant and finite dimensional control systems.
This notion of output controllability has been first introduced in the 60s by Bertam and Sarachik
and their results have been extended some years later by Kreindler and Sarachik. More precisely,
they extend the well-known necessary and sufficient conditions for (state) controllability, i.e.
Kalman rank condition and positivity of the controllability Gramian, to the framework of output
controllability.
We will see in this talk that the Hautus test condition can also be adapted for output controllability.

In a second part of the talk, we will consider the notion of “long-time output controllability”.
More precisely, the question we would like to answer is: “If a desired output can be reached in
some time T>0, can we design a control such that this output remains constant for later times
t>T?” We will see that necessary and sufficient conditions for long-time output controllability
are the ones of output controllability for an extended system output.

These works have been made in collaboration with B. Danhane, M. Jungers and M. Lazar.

Darko Mitrovic,
Global Controllability for Quasilinear Non-negative Definite System of ODEs and SDEs

We consider a control problem for a system of quasi-linear ODEs and SDEs with a non-negative definite
symmetric matrix of the system. The strategy of the proof is the standard linearization of the system
by fixing the function appearing in the nonlinear part of the system, and then applying the Leray-Schauder
fixed point theorem. We shall also need the continuous induction arguments to prolong the control to
the final state which is a novel approach in the field. This enables us to obtain controllability
for arbitrarily large initial data (so called global controllability).

Ivan Veselić,
Null controllability for the semigroup of the harmonic oscillator

More than 15 years ago it was established that control on any set of positive Lebesgue measure
is sufficient to drive the solution of the free heat equation on a bounded domain with Dirichlet boundary conditions
to zero at any prescribed positive time.
More recently,  it was established that the free heat equation in the whole euclidean space
is null controllable iff the sensor set is thick.
In the first case the spatial domain is bounded and the generator has purely discrete spectrum.
In the second case the spatial domain is unbounded and the generator has purely continuous spectrum.
Our interest is to reconcile and interpolate these two phenomena.
This can be done, on one hand, by a quantitative analysis of the
control cost estimates and their dependence on the geometric features
of the spatial domain and the sensor set.
On the other hand, the control problem for the semigroup generated by the harmonic
oscillator exhibits a mix of the phenomena spelled out above: The spatial domain is unbounded,
but the generator has purely discrete spectrum.
We present new uncertainty principles of Hermite functions
that imply null controllability for sensor sets that are much sparser than thick sets.
(This is joint work with A.~Dicke and A.~Seelmann).

Enrique Zuazua,
Dynamics and control of a guiding model

We model, simulate and control the guiding problem for a herd of evaders under the action of repulsive drivers. The problem is formulated in an optimal control framework, where the drivers (controls) aim to guide the evaders (states) to a desired region of the Euclidean space.
Classical control methods allow to build coordinated strategies so that the drivers successfully drive the evaders to the desired final destination.
But the computational cost quickly becomes unfeasible when the number of interacting agents is large.
We present a method that combines the Random Batch Method (RBM) and Model Predictive Control (MPC) to significantly reduce the computational cost without compromising the efficiency of the control strategy.
An analysis of the comnvergence of these methods, together with some open problems will also be presented.

This talk is based on joint work with Dongnam Ko, from the Catholic University of Korea, and Daniël Veldman (FAU).

Registration

Live participation: 100 EUR -750 HRK (VAT INCLUDED)

Online participation: 70 EUR – 500 HRK (VAT INCLUDED)

INSTRUCTIONS FOR FEE PAYMENT
Details of bank account:

*Beneficiary: Sveučilište u Dubrovniku:
Address: Branitelja Dubrovnika 29, Dubrovnik, Croatia
Telephone number: +385 20 445 925
Fax: +385 20 435 590

Bank name: PRIVREDNA BANKA ZAGREB, CROATIA

Branch address: PBZ Zagreb
Radnička cesta 50
Tel: + 385 01 6360000
Fax: +385 01 6360060

Payments (in €) to be made to the following account number:

Account No: IBAN: HR08 2340 0091 1101 3501 5
Call to account number: 31-20-144
SWIFT Number: PB ZG HR 2X
VAT number HR01338491514
*we kindly ask you to insert in your payment description: ConDyS Workshop 2021.

Payments (in HRK) to be made to the following:

Account number: IBAN: HR08 2340 0091 1101 3501 5 (Privredna banka Zagreb)
Call to account number: 31-20-144
*we kindly ask you to insert in your payment description: ConDyS Workshop 2021.

Please note that a confirmation of your payment (a copy of your Conference Fee payment slip)
is required. You are kindly requested to submit this copy to us in scanned form by
e-mail to: