1: \begin{definition}\label{def:Improved Outer Bound-vector}

2: %Consider a set of correlated sources $Y_s, s\in {\mathcal S}$ with underlying probability distribution $P_{Y_{\mathcal S}}(\cdot)$. From this distribution, construct binary partition random variables

3: %$A_{\langle \alpha \rangle}$ as described in Theorem \ref{thm2}. 

4: %%

5: %Let $\mathcal R''(\overline{\Gamma^*})$ be the set of all link capacity tuples $\mathbf{{C}} = (C_{e}: e \in \mathcal E)$ such that there exists an almost entropic function $ h  \in \overline{\Gamma^*}$ satisfying the constraints

6: %%For any

7: %%for given source random

8: %%variables $Y_{\mathcal S} \sim P_{Y_{\mathcal S}}(\cdot)$ and auxiliary random variables $K_{\mathcal L} \sim P_{K_{\mathcal L}|Y_{\mathcal S}}(\cdot)$ which supports the following constraints for all

9: %\begin{align}

10: %h (X_{\mathcal W}, B_{\langle \alpha \rangle})-H(Y_{\mathcal W},A_{\langle \alpha \rangle})&=0\\

11: %%&\forall \ \mathcal W \subseteq \mathcal S, \langle \alpha \rangle \in \Omega  \\

12: % h (U_{e}|\{Y_{s}: a(s) \rightarrow e\},\{U_{f}: f \rightarrow e\}) &= 0\\

13: % h (Y_{s}:u \in b(s)| Y_{s'}: s' \rightarrow u, U_{e}: e \rightarrow u) &=0\\

14: %% &\forall u \in b(s), s \in \mathcal S\\

15: % C_{e} -h (U_{e}) &\geq 0

16: %%H(Y_{\mathcal W},A_{\langle \alpha \rangle})&= h (X_{\mathcal W}, B_{\langle \alpha \rangle})\\%,\forall \ \mathcal W \subseteq \mathcal S, \langle \alpha \rangle \in \Omega  \\

17: %% 0&=h (U_{e}|Y_{s}: a(s) \rightarrow e,U_{f}: f \rightarrow e) \\

18: %% 0&=h (Y_{s}: u \in b(s)| Y_{s'}: s' \rightarrow u, U_{e}: e \rightarrow u), \nonumber\\ %&\text{ \ \ \ \ \ \ \ \ \ \ \ \  \ \ \ \ \ \ \ \ \ \ \ \ \ \ } \forall u \in b(s)\\

19: %% C_{e} &\geq h (U_{e})

20: %\end{align}

21: %for all $\mathcal W \subseteq \mathcal S, \langle \alpha \rangle \in \Omega, e\in \mathcal E, u \in b(s)$ and $s\in\mathcal S$.

22: %\end{definition}

23: