solve $\mu\cdot P^6=d$ with $P$ a matrix

39 Views Asked by At

For this problem I'm asking for references or any toolbox on python that can help me, since I'm absolutly lost, even how to search some help on internet.

Consider $P$ a $16\times 16$ matrix, wich is know except by $15$ parameters. Consider some row verctor $\mu$ that in my case is $\begin{pmatrix}1& 0 &0 &0 &... &0\end{pmatrix}$, and some target vector $d\in \mathbb{R}^{16}$. The goal is to determine the $15$ parameters such that $\mu P^6=d$

The context is that I have a Markov Chain $(M_i)_{i=0}^{6}$ with 16 states, where the transition matrix $P$ have some parameters, and I'd like to find those parameters in order to get some given distribution on time 6.

Thank you.

1

There are 1 best solutions below

0
On

$\mu P^6$ is the first row of $P^6$, that is the probability distribution at time $6$ when we come from the state $1$ at time $0$. Necessarily $(*)$ "your $15$ unknown entries must explicitly appear in this first row"; else you cannot determine them! If $(*)$ is true, then you have to solve a system of $16-1=15$ equations (because the sum of any row is $1$) in $15$ unknowns. If you are lucky, then the degrees of your equations may be low ($1$ for example if you are very lucky).