If I have four states in
\ket{S_1}=\frac{1}{\sqrt{2}}(\ket{00}+\ket{+1})_{AB}
\ket{S_2}=\frac{1}{\sqrt{2}}(\ket{-1}-\ket{10})_{AB}
\ket{S_3}=\frac{1}{\sqrt{2}}(\ket{00}+\ket{+1})_{AB}
\ket{S_4}=\frac{1}{\sqrt{2}}(\ket{00}+\ket{-1})_{AB},
and its density matrix is
\rho=frac{1}{2}(\ket{S_1}\bra{S_1}+\ket{S_2}\bra{S_2}+\ket{S_3}\bra{S_3}+\ket{S_4}\bra{S_4}).
Using the Holevo's theorem the bound of mutual information can be calculated as
I(X;Y)\leqslant S(\rho)-\frac{1}{4}(S(\ket{S_1}\bra{S_1})+S(\ket{S_2}\bra{S_2})+S(\ket{S_3}\bra{S_3})+S(\ket{S_4}\bra{S_4}))=1.60087603669285.
How to calculate the best measurement probability p from the accessible information?
For example to simple explain my question, a density matrix in mixed state
\rho=\frac{1}{2}(\ket{0}\bra{0}+\ket{+}\ket{+}),
where
\ket{0}=\begin{pmatrix}1 \\ 0\end{pmatrix} and \ket{+}=\frac{1}{\sqrt{2}}\begin{pmatrix}1 \\ 1\end{pmatrix}.
We can then calculate the mutual information and its accessible information.
After that, we can estimate the best p from Shannon entropy as
-p\log_2(p)-(1-p)\log_2(1-p)=S(\rho)-\frac{1}{2}(S(\ket{0}\bra{0})+S(\ket{+}\bra{+}))=0.600876036692856,
and thus
p=0.85355.
I am a beginner. If the concept is wrong, please correct me.