MATLAB Library for Trace Fitting


KPC-Toolbox is a library of MATLAB functions for fitting an empirical dataset into a Markov model such as a phase-type distribution (PH) or a Markovian arrival process (MAP). The user provides the input trace describing, e.g., packet inter-arrival times, job response times, job service times. Based on the statistical descriptors of the trace, the KPC-Toolbox automatically fits the dataset into a PH or a MAP using exact and approximate moment matching techniques. PHs and MAPs can be easily integrated in queueing models for performance prediction and are compositional with Markov models in general. PHs include hyper-exponential, Erlang, hypo-exponential/generalized Erlang, acyclic phase-type distributions as special cases. Special cases of MAPs are instead the interrupted Poisson process (ON/OFF), the switched Poisson process, and the Markov Modulated Poisson process (MMPP).

The KPC-Toolbox is distributed to the public for research and testing. If you have any questions, ideas or suggestions, please contact the maintainer: g DOT casale AT imperial DOT ac DOT uk.

This work has been supported by the National Science Foundation under grants ITR-0428330, CNS-0720699, and CCF-0811417.

Latest News

MAM7 Demonstration

2011 June

The latest release of the KPC-Toolbox (version 0.3.0) will be demonostrated at the Seventh International Conference on Matrix-Analytic Methods in Stochastic Models (MAM7) June 13-16, 2011 - Columbia University, New York, NY, USA

QEST Best Student Paper Award!

2008 September

The paper accepted to International Conference on Quantitative Evaluation of Systems (QEST) 2008 has received the QEST Best Student Paper Award . It can be found here.

QEST paper submitted and appendix uploaded

2008 April

A paper submitted to QEST for the tool of KPC. The Appendix is uploaded and can be viewed online. See the category of "Publications"

Webpage Online

2008 March

The KPC Toolbox webpage is being constructed. Two Papers for fundamentals of KPC have been uploaded. Click the link under Papers on the left panel: Fundamentals of KPC, you will find two published papers downloadable. These two papers explained how KPC(Kronecker Product Composition) works for process generation

Webdesign service by Sarkis. Outsourcing by FreelanceWebmarket.