Equivalences between Markov Renewal Processes.

Abstract

We define a form of equivalence between Markov-renewal processes that includes strong and weak lumpability as special cases, and examine its properties. If (X sub n, T sub n) is a Markov-renewal process with kernel Q(t) and (Z sub n, S sub n) is a Markov-renewal process with kernel Y(t), then it is shown that (X sub n, T sub n) and (Z sub n, S sub n) are equivalent if and only if there is a certain homomorphism between the matrix rings generated by Q(t), t an element from zero thru infinity and Y(t), t an element from zero thru infinity. The equivalence is identical to weak lumpability in the case where (Z sub n, S sub n) is a renewal process. Although the conditions for strong lumpability can be written in an attractive form, they are too restrictive to be of any real interest. Weak lumpability is of more interest since (as will be shown) it occurs in less trivial examples, but the necessary conditions are very complicated. The equivalence defined herein has the advantage of having simple necessary and sufficient conditions. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1978
Accession Number
ADA064544

Entities

People

  • Burton Simon

Organizations

  • University of Michigan

Tags

DTIC Thesaurus Topics

  • Air Force
  • Computations
  • Computers
  • Contracts
  • Control Systems Engineering
  • Distribution Functions
  • Eigenvalues
  • Eigenvectors
  • Engineering
  • Markov Chains
  • Michigan
  • Military Research
  • Probability
  • Probability Distributions
  • Random Variables
  • Scientific Research
  • Steady State

Fields of Study

  • Mathematics

Readers

  • Calculus or Mathematical Analysis
  • Graph Algorithms and Convex Optimization.
  • Ocean-Atmosphere Mesoscale Modeling, Data Assimilation, and Flux Boundary Layers