IoT 15(4): e2

Research Article

A heuristic procedure for compact Markov representation of PH distributions

Download289 downloads
  • @ARTICLE{10.4108/icst.valuetools.2014.258174,
        author={Illes Horvath and Miklos Telek},
        title={A heuristic procedure for compact Markov representation of PH distributions},
        journal={EAI Endorsed Transactions on Internet of Things},
        volume={1},
        number={4},
        publisher={EAI},
        journal_a={IOT},
        year={2015},
        month={2},
        keywords={ph distribution, minimal representation, feedback erlang block, unicyclic block},
        doi={10.4108/icst.valuetools.2014.258174}
    }
    
  • Illes Horvath
    Miklos Telek
    Year: 2015
    A heuristic procedure for compact Markov representation of PH distributions
    IOT
    EAI
    DOI: 10.4108/icst.valuetools.2014.258174
Illes Horvath1,*, Miklos Telek2
  • 1: MTA-BME Information Systems Research Group
  • 2: Department of Networked Systems and Services Budapest University of Technology and Economics
*Contact email: horvath.illes.antal@gmail.com

Abstract

The minimal Markovian representation of PH distributions is an open research problem, which was actively investigated during the last two decades. We present a numerical method for finding small Markovian representation of PH distributions and investigate the general quality of the method by comparing the size of the obtained representation with the size of the representation obtained by alternative methods. Our numerical method intends to find a small Markovian representation. We report examples when the obtained representation is larger than the minimal Markovian representation.