For a finite Markov chain whose transition rows sum to one, when is the uniform distribution stationary?
Quant ResearchLevel: IntermediateUniform stationary law
For a finite Markov chain whose transition rows sum to one, when is the uniform distribution stationary?
Hints
- Multiply the uniform row vector by P.
- Distinguish doubly stochastic from symmetric.
Answer
Exactly when every column also sums to one. The j-th coordinate of uniform*pi P is (1/n)sum_i Pij and must equal 1/n. Symmetric matrices satisfy this, but symmetry is not necessary; a cyclic permutation matrix is a counterexample.
Source and licence: Lehman, Leighton & Meyer — Mathematics for Computer Science (2018-06-06) · CC-BY-SA-3.0 · answer: Quantoros editorial