Transition rate matrix

From HandWiki
Short description: Matrix describing continuous-time Markov chains

In probability theory, a transition rate matrix (also known as an intensity matrix[1][2] or infinitesimal generator matrix[3]) is an array of numbers describing the instantaneous rate at which a continuous time Markov chain transitions between states.

In a transition rate matrix Q (sometimes written A[4]) element qij (for i ≠ j) denotes the rate departing from i and arriving in state j. Diagonal elements qii are defined such that

[math]\displaystyle{ q_{ii} = -\sum_{j\neq i} q_{ij}. }[/math]

and therefore the rows of the matrix sum to zero (see condition 3 in the definition section).

Definition

A transition rate matrix [math]\displaystyle{ Q }[/math] satisfies the following conditions[5]

  1. [math]\displaystyle{ 0 \leq -q_{ii} \lt \infty }[/math]
  2. [math]\displaystyle{ 0 \leq q_{ij} : \mathrm{for}\; i \neq j }[/math]
  3. [math]\displaystyle{ \sum_j q_{ij} = 0 : \mathrm{for}\;\mathrm{all}\; i }[/math]

Up to a global sign, a large class of examples of such matrices is provided by the Laplacian of a directed, weighted graph. The vertices of the graph correspond to the Markov chain's states.

Properties

The transition rate matrix has following properties:[6]

  • There is at least one eigenvector with a vanishing eigenvalue, exactly one if the graph of [math]\displaystyle{ Q }[/math] is strongly connected.
  • All other eigenvalues [math]\displaystyle{ \lambda }[/math] fulfill [math]\displaystyle{ 0 \gt \mathrm{Re}\{\lambda\} \geq 2 \min_i q_{ii} }[/math].
  • All eigenvectors [math]\displaystyle{ v }[/math] with a non-zero eigenvalue fulfill [math]\displaystyle{ \sum_{i}v_{i} = 0 }[/math].

Example

An M/M/1 queue, a model which counts the number of jobs in a queueing system with arrivals at rate λ and services at rate μ, has transition rate matrix

[math]\displaystyle{ Q=\begin{pmatrix} -\lambda & \lambda \\ \mu & -(\mu+\lambda) & \lambda \\ &\mu & -(\mu+\lambda) & \lambda \\ &&\mu & -(\mu+\lambda) & \lambda &\\ &&&&\ddots \end{pmatrix}. }[/math]

See also

References

  1. Syski, R. (1992). Passage Times for Markov Chains. IOS Press. doi:10.3233/978-1-60750-950-9-i. ISBN 90-5199-060-X. 
  2. Asmussen, S. R. (2003). "Markov Jump Processes". Applied Probability and Queues. Stochastic Modelling and Applied Probability. 51. pp. 39–59. doi:10.1007/0-387-21525-5_2. ISBN 978-0-387-00211-8. 
  3. Trivedi, K. S.; Kulkarni, V. G. (1993). "FSPNs: Fluid stochastic Petri nets". Application and Theory of Petri Nets 1993. Lecture Notes in Computer Science. 691. pp. 24. doi:10.1007/3-540-56863-8_38. ISBN 978-3-540-56863-6. 
  4. Rubino, Gerardo; Sericola, Bruno (1989). "Sojourn Times in Finite Markov Processes". Journal of Applied Probability (Applied Probability Trust) 26 (4): 744–756. doi:10.2307/3214379. https://hal.inria.fr/inria-00075739/file/RR-0812.pdf. 
  5. Norris, J. R. (1997). Markov Chains. doi:10.1017/CBO9780511810633. ISBN 9780511810633. 
  6. Keizer, Joel (1972-11-01). "On the solutions and the steady states of a master equation" (in en). Journal of Statistical Physics 6 (2): 67–72. doi:10.1007/BF01023679. ISSN 1572-9613. Bibcode1972JSP.....6...67K. https://doi.org/10.1007/BF01023679.