Necessary and sufficient conditions for a transition matrix to have a limit

65 Views Asked by At

Let $T$ be an $n \times n$ transition matrix, i.e. the rows sum to 1 and the entries all lie in the interval $[0,1]$. What are necessary and sufficient conditions for which the limit $lim_{n \rightarrow \infty} T^{n}$ exists?

I recall reading somewhere (but don't remember where) that this is true if and only if $1$ is an eigenvalue of $T$ with geometric multiplicity 1, and all other eigenvalues have length strictly less than 1. But this does not seem to be true, for the following matrix:

    [0, 0.5, 0.5 , 0   , 0   , 0  , 0  ],
    [0, 0  , 0.33, 0.33, 0.33, 0  , 0  ],
    [0, 0  , 0   , 1   , 0   , 0  , 0  ],
    [0, 0  , 0   , 0   , 0.5 , 0.5, 0  ],
    [0, 0  , 0   , 0   , 0   , 0  , 1  ],
    [0, 0  , 0.5 , 0   , 0   , 0  , 0.5],
    [0, 0  , 0   , 1   , 0   , 0  , 0  ]

According to MATLAB, the limit of its powers appears to converge to:

         0         0    0.1663    0.1650    0.0825    0.0825    0.4988
         0         0    0.0825    0.3300    0.1650    0.1650    0.2475
         0         0         0    1.0000         0         0         0
         0         0         0         0    0.5000    0.5000         0
         0         0    0.2500         0         0         0    0.7500
         0         0    0.2500         0         0         0    0.7500
         0         0         0    1.0000         0         0         0

Yet its eigenvalues are $1,0,-0.5000 + 0.8660i,-0.5000 - 0.8660i$ the last two of which have length strictly equal to 1.