Skip to content

Main topic: Agent-based models

Some systems analyzed by Physics of Risk are very complex and are impossible to understand by using traditional for physicist – reductionist – approach. Such systems are not reductive, because of importance of interactions between different parts of systems and the parts themselves. There is one example which should be known to everybody – human body. By cutting human heart out or by tinkering with human brain we won’t be able to clearly understand how human organism works! In order to do so we must study all internal organs and their interrelations.

Though human body is at the same time perfect and slightly inapt example of complex system. It is so, because that each internal organ has it’s own differing function and thus differing structure. Individuals within financial markets in this sense are different. At least from the bird’s eye view they appear to be identical as they all perform similar function while participating in global exchange of money and goods. It is thought that essential properties of time series from financial markets could be explained by studying complex interactions of somewhat similar and somewhat different individuals within financial markets.

Aforementioned possibility is explored through the so-called agent based models. These models replace individuals from actual systems, not necessarily limited to financial markets, with abstract objects, so-called agents, which base their behavior on simple rules. In this section we will show that, despite simplifications in behavior of individuals, complex behavior of model system can be obtained using agent based modeling.

More models »

Prisonner’s dilemma

Previously we wrote about mathematical “puzzle” originating from a TV game (see the description of the Monty Hall problem). This time we shall consider the opposite case – the mathematical “game” used as a base for a TV game. Watch a fragment of the “Golden Balls” final stage called “Split or steal”.

The game is very simple, yet it possesses no correct solution or optimal strategy. Interestingly enough it can also be used as model for understanding social behavior of humans.
Žaidimo esmė yra elementari, bet kaip nebūtų keista – teisingo sprendimo ar laiminčios strategijos čia nėra ir negali būti. Įdomu ir tai, kad toks elementarus modelis gali būti panaudotas žmonių visuomeniškumo ir egoistiškumo supratimui 1.

RCL: Call for applications for Students’ Research Practice in Lithuania in summer 2013

svietimas

Research Council of Lithuania has announced a call for applications for students’ research practice in Lithuania in summer of 2013. The contributors towards Physics of Risk website, dr. (HP) Vygintas Gontis and PhD student Aleksejus Kononovičius, offer two topics for the research practice. The offered topics are mainly based on the following topics, previously published on this website:

More information is available from the Research Council of Lithuania website (see the call here) and from the portal dedicated to the project “Promotion of Students’ Scientific Activities” (see here).

FuturICT: participatory science and computing for our complex world

futurict

The most active and well known scientists contributing towards FuturICT project have organized their ideas into few articles, which discuss the current situation in contemporary complex social system modeling. These articles we published in a special issue of “The European Physical Journal Special Topics”. Most of the articles are available for free, thus we invite you to familiarize yourselves with them.

Read the special issue of EPJ »

Lithuanian FuturICT Hub »

Vygintas Gontis: Econophysics – brand new outlook into social sciences

Gontis_MKS_01

Econophysics is not a very well known interdisciplinary field of research in Lithuania. There are very few Lithuania based scientists who are related to this very young and rapidly developing branch of Physics. Therefore we, being ones of the few, feel an obligation to introduce Econophysics to the other scientists, businessmen and the whole broader society. In this text we will discuss the basic ideas, current trends and the most common problems relevant to the field of Econophysics.

The name of Econophysics implies the application of physics, mostly the methods taken from Statistical Physics, to the Economics. Though until recently most of the econophysicists used to focus on the financial markets. Currently the field is expanding very rapidly and now concerns something more than just regular Economics or Finance – many physicists now attempt to tackle problems in the other social sciences. For these broader approaches there are other overlapping terms including Sociophysics, Physics of Risk, Physics of Socio-Economic Systems and other.

At this point we would like to stop the discussion on the use of the terminology. Recall that the aim of this text is somewhat different from that. We aim to familiarize you with the Econophysics! Continue reading “Vygintas Gontis: Econophysics – brand new outlook into social sciences” »

Seminar at VU MIF: Agent-based Versus Macroscopic Modeling of Competition and Business Processes in Economics and Finance

Gontis_MKS_01

Topic: “Agent-based Versus Macroscopic Modeling of Competition and Business Processes in Economics and Finance”
Speaker: dr. Vygintas Gontis
Briefly: The talk will be focused on the agent-based and stochastic modelling done by the Department of the Theory of Processes and Structures of the VU ITPA.
When? 6th of November, 17:00.
Where? VU Faculty of Mathematics and Informatics (Naugarduko g. 24, Vilnius), 400 auditorium.
Organized by: Department of the Mathematical Analysis of the VU MIF.

Seminar at VU Faculty of Physics: Brief introduction into the Physics of Risk

kononovicius

Topic: “Physics is not a risk: Brief introduction into the Physics of Risk”
Speaker: Aleksejus Kononovičius
Briefly: Social sciences have accomplished many different things. Yet it should be evident that there is a place for improvements – to look into the old and new social problems a bit differently. As many of the social problems are strongly non-linear and very complex, the physicists’ point of view is very useful. This, new, point of view is known as Physics of Risk.
When? 18th of October, 17:00.
Where? VU Faculty of Physics (Saulėtekio al. 9, III rūmai, Vilnius), 201 auditorium.
Organized by: VU Faculty of Physics Students Scientific Association.
Facebook event: here.

Slides: download (in Lithuanian).

Agent based prey-predator model

lv-agent

The simplest ecological system can be constructed from the two interacting species, ex. prey and predator. This kind of system is very interesting in the terms of Physics of Risk primarily because it is nonlinear 1, and due to being real life example of competition (conflict). Also there are few known simple models for the prey-predator interaction. Among them there are both macroscopic, Lotka-Volterra equations, and microscopic, agent-based, models. In this text we continue the previous discussion by considering the agent-based model. Continue reading “Agent based prey-predator model” »

EURO 2012 starts next week

euro-2012

EURO 2012 starts next week in Vilnius! EURO 2012, European Conference on Operational Research, is huge scientific event popular both among scientists and businessmen. Operational research tackles very challenging, and thus scientifically interesting, and highly applicable topics. Some of these topics overlap with the ones we discuss here on Physics of Risk – financial market modelling, risk management and analysis, decision making. Many reports will also consider logistics, optimization and network analysis.

On of our authors will join numerous scientists from Lithuania and abroad giving talks in the conference. On Monday, the 9th of July, he will give a talk on “Herding behavior of agents as a background of financial fluctuations”, Our works on the applications of Kirman model will be the key point of the talk.

Read more on the conference website: euro-2012.lt.

Simple rules and the complexity of collective behavior

1 pav. Iliustracija kaip elementariausias taisyklių pakeitimas sąlygoja elgesio pobūdžio pokytį agentų modelyje.

It should be pretty straightforward why the agent based models appear to be very simple. They do poses technical simplicity, which my cause some doubts on the actual usefulness of the agent based models. But the fact is that they are able to reproduce complex, non-trivial behavioral patterns. Continue reading “Simple rules and the complexity of collective behavior” »

K. Staliunas on econophysics

staliunas

Traditionally it is thought that economics and physics are very different irreconcilable sciences. Physicists sometimes claim that economists are aware only of addition and subtraction, while only very rarely they attempt to use multiplication or even division. Economists on the other hand think that physicists are busy dividing molecules into atoms, create atomic weapons and lasers and are able to fix AC sockets, while having absolutely no understanding in other, nontechnical, fields. Evidently this oversimplified thinking is incorrect – economics and physics have things in many common. And with the time more common things are discovered.

Continue reading “K. Staliunas on econophysics” »