Matrix stochastic game with Q-learning

Abstract

The model of matrix stochastic game for decision-making in the conditions of uncertainty is developed. The method of Q-learning for stochastic game solving with a priori unknown gains matrices is offered. The formulation of a game problem is executed. The Markovian recurrent method and algorithm for the game solving are described. Results of computer modelling of stochastic game with Q-learning are received and analysed.

Authors and Affiliations

Petro Kravets

Keywords

Related Articles

Approximation of trustworthiness of information objects of subject area ontology based on polynomial splines

The method of coefficient approximation of information objects domain ontologies reliability is proposed in the article. This method is based on polynomial splines. It makes it possible to remove unnecessary ontology obj...

Forecasting network requests numbers to cloud application

The article contains analysis of existing time series forecast methods. It is estimated how these methods fit to time series of network requests to a cloud application. Optimal forecasting methods have been chosen for di...

Intelligent Analysis of Data Structures and Mathematical Software of Formulas Editor

This article is about determination of process of adaptation, data structure of algorithms editor. Mathematical model of data structure of algorithms editor, hierarchical-modular construction principle of editor, of adap...

Modeling of energy potential of wind power plant

This paper deals with the research and comparative analysis of methods of determining the probability of wind speed’s recurrence by using Weibull distribution, gamma distribution and lognormal distribution based on stati...

Methods of Knowledge Representation and Processing in Systems With Situational Awareness

In the article, the methods of representation and processing of knowledge in systems with situational awareness are described. The definitions of concepts and mathematical formalization of situation and situational aware...

Download PDF file
  • EP ID EP617419
  • DOI -
  • Views 165
  • Downloads 0

How To Cite

Petro Kravets (2015). Matrix stochastic game with Q-learning. Vìsnik Nacìonalʹnogo unìversitetu "Lʹvìvsʹka polìtehnìka". Serìâ Ìnformacìjnì sistemi ta merežì, 814(), 71-80. https://europub.co.uk/articles/-A-617419