Selected Decomposition Methods with Applications to Large Scale Systems

Organizer: Mathematical Sciences Research Unit
Credits:  2
Dates: September 20-25, 2017
Place University of Oulu, Linnanmaa Campus


The main goal of the lecture series is to consider efficient solution methods of the problems related to intelligent data analysis, whose efficient solution appears crucial for various fields of applications, such as machine learning, signal, speech and image recognition and processing, and many others. These problems are usually formulated as rather simple optimization problems, but dealing with huge sets of inexact, incomplete, and spatially distributed data, which imposes essential restrictions on both the basic mathematical models and choice of the methods. The lecturer is Prof. Igor Konnov from the Kazan Federal University, Kazan, Russia.

Content of the lectures

Wed 20 at 9-11 in Aspire, Tellus: Elements of Non Differentiable Optimization (Non differentiable convex functions, Subgradient methods)

Wed 20 at 14-16 in Lecture Room PR 102: Lagrangian Duality (Lagrangian dual methods, Application to decomposable problems)

Thu 21 at 12-14 in Lecture Room IT134: Regularization Methods (Convex and non-convex optimization, iterative regularization)

Fri 22 at 12-14 in Lecture Room IT134: Applications to Market Models (Centralized and decentralized models, game models)

Mon 25 at 10-12 in Lecture Room PR 126A: Solution Methods for Limit Problems (Limit problems, penalty methods)


Description of lecturer

D.Sc. Igor Konnov (  is a professor at the Department of System Analysis and Information Technologies, Kazan Federal University, Kazan, Russia. His research topics include theory, methods, and applications of nonsmooth optimization, equilibrium problems, and variational inequalities. He has published five books and more than 230 peer reviewed scientific papers in these fields.


Registrations in the (optional)

More information: Prof Erkki Laitinen (erkki.laitinen at



Last updated: 15.9.2017