I'm having a hard time reading the Bruce Kitchen's book "Symbolic Dynamics-One-sided, Two-sided and Countable State Markov Shifts". In page 16, he introduces the Perron-Frobenius Theory with some basic knowledge, and I cannot understand them.
Here are the definitions:
A real, nonnegative, square matrix $A$ is called irreducible if for every pair of indices $i$ and $j$ there is an $l>0$ with $(A^l)_{ij}>0$. This is equivalent to saying that the related graph $G_A$ is strongly connected.
With the same condition on the matrix $A$, given an index $i$, let $p(i)=gcd\left\{ l:(A^l)_{ii}>0\right\}$. This is the period of the index i.
When A is irreducible the period of every index is the same and is called the period of A.
Then I have 2 questions on these.
Is there any motivation for using gcd to be included in the definition of the period of the index? I found from Wikipedia that this definition is equivalent to the greatest common divisor of all lengths of the possible cycles at $i$. But why do we need gcd instead of the least length of the cycle?
In the third paragraph of the definition box, I cannot found the reason for all periods of indices to be same for the irreducible matrix $A$. How can I prove this?