ICM Joint Academic Workshop on Stochastic and Numerical Methods

Text

Study location Gamma (09:15-12:00) and U2-016 (13:15-17:00)

2026-06-02 09:00–17:00

ICM Joint Academic Workshop on Stochastic and Numerical Methods

Date: 2026-06-02

Location: Gamma (09:15-12:00) and U2-016 (13:15-17:00), campus Västerås, Mälardalen University (MDU)

About the workshop

Mälardalen University, in collaboration with National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute” and Taras Shevchenko National University of Kyiv, is pleased to announce the ICM Joint Academic Workshop on Stochastic and Numerical Methods.

The workshop brings together researchers in stochastic analysis, numerical methods, mathematical modeling, optimization, statistics, and machine learning. The aim is to strengthen international academic cooperation and create opportunities for future joint research and educational initiatives between the participating institutions.

The workshop will provide a platform for presenting recent research results, promoting interdisciplinary discussions, and fostering long-term academic collaboration between Sweden and Ukraine.

Organizing institutions

(host) Mälardalen University (MDU), Faculty of Philosophy, Department of Business and Mathematics, Västerås, Sweden

National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Kyiv, Ukraine

National Taras Shevchenko University, Kyiv, Ukraine

Programme

Morning Session

09:15–09:30 — Registration & Mingle

09:30–10:00

Speaker: Oleg Klesov

Affiliation: National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”

Title: A characterization of the normal distribution

Abstract: According to a Gauss theorem, if the MLE for the unknown population mean equals the sample mean for each sample of an arbitrary size, then the distribution of the population is normal. We prove this result under the assumption that MLE equals the sample mean for each sample of a fixed size >2.

10:00–10:30

Speaker: Iryna Bodnarchuk

Affiliation: Taras Shevchenko National University of Kyiv, Ukraine

Title: Analysis of entropies and divergences in probability models

Abstract: This is a joint work with Yuliya Mishura and Kostiantyn Ralchenko. We establish the properties of the Kullback–Leibler divergence and the Shannon, modified Shannon, Rényi, generalized Rényi (with one and two parameters), Tsallis, and Sharma–Mittal entropies for such distributions as the exponential, gamma, chi-squared, Laplace, and log-normal ones.

10:30–11:00

Speaker: Rostyslav Yamnenko

Affiliation: Taras Shevchenko National University of Kyiv, Ukraine

Title: Sub-Gaussian γ-reflected process of fractional Brownian motion

Abstract: The talk focuses on γ-reflected processes driven by sub-Gaussian self-similar inputs

Wγ(t)=X(t)−ct−γ inf⁡s≤t(X(s)−cs), c>0.

Such processes arise in risk theory in models with taxation based on the loss-carry-forward scheme, where a proportion γ of incoming premiums is paid as tax. Estimates for the distribution of the supremum are proposed. Special attention is devoted to the case where the input process is a fractional Brownian motion.

11:00–11:15 — Coffee Break

11:15–11:45

Speaker: Iryna Holichenko

Affiliation: National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”

Title: Interpolation of periodically correlated sequences with missing observations

Abstract: The problem of the optimal linear estimation of a linear functional depending on the unknown values of periodically correlated stochastic sequence is considered. The estimate is based on observations of the sequence with noise. Formulas for calculating the mean square error and the spectral characteristic of the optimal linear estimate of the functional are proposed in the case where spectral densities are exactly known. Formulas that determine the least favorable spectral densities and the minimax spectral characteristics are proposed in the case of spectral uncertainty, where the spectral densities are not exactly known while some class of admissible spectral densities is given.

11:45–12:15

Speaker: Iryna Alieksieieva

Affiliation: National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”

Title: On the Relation between Maxmin and Nash Equilibrium in Bimatrix Games

Abstract: The talk is devoted to the study of the relation between maxmin strategies and Nash equilibrium in two-person bimatrix games. For certain classes of payoff matrices, sufficient conditions are established under which a player’s guaranteed payoff under a maxmin strategy coincides with the payoff attained at a Nash equilibrium. Classical examples illustrating essentially different cases of such coincidence are also discussed.

12:15–13:30 — Lunch and Discussions

Afternoon Session

13:30–14:00

Speaker(s): Vitaliy Miroshnychenko, Tetiana Skorobohach

Affiliation: Taras Shevchenko National University of Kyiv, Ukraine, National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute

Title: Differentiable upper bound objectives for statistical learning problems for mixture of varying concentrations

Abstract: The talk focuses on studying mixture models with varying concentrations (MVC). Existing MVC approaches assume either fixed mixing probabilities or parametric models estimated from auxiliary data external to the regression task. We propose a method that eliminates the need for auxiliary data. Direct optimization with MVC minimax weights is challenging because the weights may take negative values. To address this issue, we replace the original objective with a differentiable upper bound. The proposed loss function provides a trade-off between component balancing and prediction accuracy. Simulation experiments on Gaussian regression mixtures demonstrate improved regression parameter estimates convergence compared with classical Gaussian mixture regression estimators.

14:00–14:30

Speaker: Vadym Ponomarov

Affiliation: Taras Shevchenko National University of Kyiv, Ukraine

Title: On multicriterial optimization of retrial queueing systems

Abstract. This presentation addresses the problem of optimal control in retrial queueing systems. From a practical standpoint, these systems constitute a crucial class of models because they explicitly account for the phenomenon of retrial calls, thereby allowing for more precise modeling of service processes. However, the presence of a secondary flow of repeated calls complicates steady-state analysis, leaving a general analytical solution elusive. When managing real-world systems, the most critical challenges often involve optimal control and the maximization of total system income. In this work, we formulate these optimization problems in a general form and describe solution methodologies applicable to a wide range of retrial queueing architectures.

14:30–15:00

Speaker: Oksana Lahoda, Viktor Selivanov

Affiliation: National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”

Title: Special Cases of Limit Theorem for Solutions of Abstract Stochastic Difference Equations

Abstract: For solutions of an abstract stochastic equation on a half-axis, the limiting behavior of the sums of the equation’s solution elements is investigated. For special cases of the limiting behavior of the input sequence components, similar behavior for the solution is established.

15:00–15:30 — Coffee Break

15:30–16:00

Speaker: Prof. em. Dmitrii Silvestrov

Affiliation: Stockholm University/Mälardalen University, Sweden

Title: Coupling, Stochastic Lyapunov Functions, Artificial Regeneration and Ergodic Theorems

Abstract: This talk informally presents the methods of coupling, stochastic Lyapunov functions, and artificial regeneration used for getting of ergodic theorems with explicit upper bounds for convergence rates for Markov chains, regenerative, and semi-Markov-type processes presented in books [1 - 2].

[1] Silvestrov, D. (2025). Coupling and Ergodic Theorems for Semi-Markov-Type Processes II: Semi-Markov Processes and Multi-Alternating Regenerative Processes with Semi-Markov Modulation. Springer, Cham, xix+560 pp.

[2] Silvestrov, D. (2025). Coupling and Ergodic Theorems for Semi-Markov-Type Processes I: Markov Chains, Renewal and Regenerative Processes. Springer, Cham, xix+611 pp.

16:00–16:30

Speaker: Prof. Sergey Korotov

Affiliation: Mälardalen University, Sweden

Title: On two hard topics in numerical analysis and modelling

Abstract: In this talk, we will present and shortly discuss two difficult problems in numerics which are not completely solved so far in our knowledge. The first one is about generation and adaptivity of simplicial meshes with controllable (geometric) properties. The second one is related to the so-called monotone matrices and their applications to numerical algorithms.

16:30–17:00

Speaker: Assoc. Prof. Achref Bachouch

Affiliation: Mälardalen University, Sweden

Title: Time discretization approximation for Generalized BDSDEs and application for Quasilinear Stochastic PDEs

Abstract: In this talk, we investigate a numerical probabilistic method to approximate the solution of a class of quasi-linear stochastic partial differential equations (SPDEs in short) . The method is based on the probabilistic interpretation of such solution within solution of generalized backward doubly stochastic differential equations (BDSDEs in short). We prove the convergence and the rate of convergence of our numerical time-discretization scheme under only Lispchitz continuous assumptions on the coefficients. The proof is based on a generalization of the result on the Zhang-path regularity of the generalized BDSDEs. Finally, numerical experiments are given in the last section.