Operational Research Stream

Prof. Boris Goldengorin
Higher School of Economics Branch in Nizhny Novgorod, Russia

Combinatorial Optimisation

In recent years, network theory has been used to analyze many large data-sets that can be represented as a graph. In particular financial networks can be used to understand the dynamics of the market and the effects of globalization. In our models, where the vertices represent stocks and the connections are defined using price correlations has been shown to follow the power-law model. Computing clique partitions, cliques, independent sets and stars have a clear practical interpretation and their analysis can be a complementary tool in financial decision making.

In this mini-course we are going to discuss results of network optimization approaches to analyze the Chinese, Russian, USA and Sweden stock markets.

Prof. Igor Kozin
Department of Economic Cybernetics, Zaporizhzhya National University, Ukraine

Principle of Symmetry and Decision-Making Problem

In many practical management problems associated with decision-making we deal with a number of selection criteria. The belonging of problem solutions to the Pareto set is taken as condition of choice in this case. But the Pareto principle is only one of the principles of rational decision choice. What other ideas can be the basis of decisions-making? One of these ideas is to use the symmetry properties, namely, invariance with respect to a particular group of transformations in choosing the "best" solution.

Sometimes this principle provides a rational solution that is almost unique.

The second question to be considered in the lectures - the concept of "beauty" of the solution. Can such an emotional thing as beauty be described by strict mathematical relationships? Perhaps the concept of symmetry and antisymmetry gives an answer to this question.

Prof. Erik Kropat
Bundeswehr University Munich, Germany

Humanitarian Operations Research ‒ Decision Support for Crisis & Disaster Operations

Recent natural and man-made disasters affected many parts of the world and resulted in an extensive loss of life, disruption of infrastructure and significant economic impacts to regional and national economies. The randomness of impacts and the urgency of response efforts are challenging the operational analyst and demand a dynamic decision-making in an uncertain and complex environment. A far-sighted and comprehensive emergency planning in the pre-disaster phase can help to alleviate the effects of a suddenly arising critical event. Methods from computational intelligence and sophisticated agent-based simulation systems allow a detailed risk analysis and an evaluation of the vulnerability of infrastructures and supply systems. Major disasters require a cross-organizational coordination of relief activities that are usually concerned with the limited availability of resources and infrastructure to address the needs of the affected population. Integrated decision support systems and Operations Research tools enable an efficient coordination of rescue and facilitate further disaster relief activities.

In this course, we will discuss how modern Operational Research can support the operational analyst in the pre-disaster planning phase as well as the post-disaster response phase. In particular, we will address selected problems in the fields of

• city evacuation planning
• vehicle routing
• facility location problems
• humanitarian supply chains
• early warning systems.

Ms. Anna Leontjeva
Faculty of Mathematics and Computer Science, Institute of Mathematical Statistics, University of Tartu, Estonia

Introduction to Time Series Analysis and Forecasting

Time series is defined as an ordered sequence of observations repeatedly obtained over time. In other words, it concerns any data with time variable. Obviously, there is a tremendous number of applications for time series (finance, physics, business monitoring, to name a few) and thus, high demand for the ‘right’ use of it.

Usually, there are two main research areas: time series analysis and time series forecasting. The first deals with an existing signal and encompasses methods for extracting valuable information. The goal of the latter is to build a statistical model to predict feature values of the same data.

The main goal of this workshop is to give some insights how to make use of this kind of data. Firstly, short introduction and basic definitions will be discussed. After that, two practical sessions will be given in order to apply some techniques on a real datasets. No prior knowledge about time series is needed. However, basics of statistics are recommended. Students will use statistical language R.

Prof. Alexander Makarenko
Institute for Applied System Analysis of NTUU KPI, Ukraine

Concepts and Mathematical Models of Consciousness

Consciousness research is one of the main directions of modern science. This rapidly developing field may cause the next scientific revolution. In the proposed lecture we will review the existing models and concepts of Consciousness as well as some prospective ideas including new concept of Anticipating Neural Networks and their relations to understanding of Consciousness.

Also we will discuss some other subjects of Operational Research that are practically connected: Neuroscience and Modelling of Socio-economical systems which include the using of Consciousness.

Detailed information

Prof. Oleg Prokopyev
Department of Industrial Engineering, University of Pittsburgh, USA

Introduction to Integer Programming

An integer program is an optimization model, in which some or all of the decision variables may only take integer values. Many practical problems in a variety of application domains (e.g., economics, transportation, logistics, bioinformatics, and health care) may be modeled as linear or nonlinear integer programs. The lecture will cover standard approaches used for formulating and solving linear as well as some classes of nonlinear integer programs. We will also overview several well-known applications of integer optimization.

Mr. Konstantin Tretyakov
Department of Computer Science, Tartu University, Estonia

Introduction to Data Mining

Data mining (which is often referred to as the machine learning, pattern analysis and frequently defined bymany other names) is one of central disciplines of modern-day computational science. Data mining deals with the problem of extracting (or "learning") "knowledge", "models", "algorithms" or other types of useful information from raw data.

The course aims to give an overview of basic data mining and machine learning techniques. We shall discuss both theoretical basis and practical implementation as well as application issues of data mining.

Detailed information

Prof. Gerhard-Wilhelm Weber
Department of Financial Mathematics and Department of Scientific Computing, Institute of Applied Mathematics, Middle East Technical University, Ankara, Turkey

Financial Mathematics

and Related Aspects of Environment, Education and Development

We approach, view and investigate the Financial Sector from its environment, to which nature (sciences), education, infrastructure and industry belong. In fact, our prospective is a one that takes into account the living conditions of the people in our countries and in the world. To these conditions this course is committed both quantitatively and qualitatively. Our course will witness this in all its parts: in motivation, in analysis, in optimization and in outlook. We are going to learn together, the further advance in scientific reasoning and creative thinking - and enjoying a pleasant and stimulation time in beautiful Kyiv that can give hope to all of us!