Algorithm 2 Iterative Algorithm for Resource Allocation.

1:Initialization: Initialize the maximum number of iterations $T_{max}$ and set iteration index $\tau =1$; and initialize the current feasible point ${\mathbf{u}}^{\left(1\right)}$, ${\mathbf{v}}^{\left(1\right)}$, ${\mathbf{p}}^{\left(1\right)}$, ${\mathbf{q}}^{\left(1\right)}$ and a penalty factor $\eta \gg 1$ 2:repeat 3:   Set the variables $\mathbf{u}$, $\mathbf{v}$, $\mathbf{p}$, and $\mathbf{q}$ to be solved by the standard convex program solvers 4:   Evaluate the convex functions $F_n^{\mathrm{A}}(\mathbf{u},\mathbf{p})$, $G_n^{\mathrm{A}}(\mathbf{u},\mathbf{p})$, $G_n^{\mathrm{B}}(\mathbf{v},\mathbf{q})$, and $H(\mathbf{u},\mathbf{v})$ 5:   According to the current point ${\mathbf{u}}^{\left(\tau \right)}$, ${\mathbf{v}}^{\left(\tau \right)}$, ${\mathbf{p}}^{\left(\tau \right)}$, and ${\mathbf{q}}^{\left(\tau \right)}$, evaluate the affine functions ${\nabla }_{{\mathbf{u}}^{\left(\tau \right)},{\mathbf{p}}^{\left(\tau \right)}}^{\mathrm{affine}}{F}_n^{\mathrm{A}}(\mathbf{u},\mathbf{p})$, ${\nabla }_{{\mathbf{u}}^{\left(\tau \right)},{\mathbf{p}}^{\left(\tau \right)}}^{\mathrm{affine}}{G}_n^{\mathrm{A}}(\mathbf{u},\mathbf{p})$, ${\nabla }_{{\mathbf{v}}^{\left(\tau \right)},{\mathbf{q}}^{\left(\tau \right)}}^{\mathrm{affine}}{F}_n^{\mathrm{B}}(\mathbf{v},\mathbf{q})$, and ${\nabla }_{{\mathbf{u}}^{\left(\tau \right)},{\mathbf{v}}^{\left(\tau \right)}}^{\mathrm{affine}}M(\mathbf{u},\mathbf{v})$ 6:   Substitute them into (48) to solve the convex problem for getting the upper-bound optimal point ${\mathbf{u}}^{\mathrm{opt}}$, ${\mathbf{v}}^{\mathrm{opt}}$, ${\mathbf{p}}^{\mathrm{opt}}$, and ${\mathbf{q}}^{\mathrm{opt}}$ 7:   Update iteration index $\tau =\tau +1$ 8:   Update the next iteration point ${\mathbf{u}}^{\left(\tau \right)}={\mathbf{u}}^{\mathrm{opt}}$, ${\mathbf{v}}^{\left(\tau \right)}={\mathbf{v}}^{\mathrm{opt}}$, ${\mathbf{p}}^{\left(\tau \right)}={\mathbf{p}}^{\mathrm{opt}}$, ${\mathbf{q}}^{\left(\tau \right)}={\mathbf{q}}^{\mathrm{opt}}$ 9:until$({\mathbf{u}}^{(\tau -1)},{\mathbf{v}}^{(\tau -1)},{\mathbf{p}}^{(\tau -1)},{\mathbf{q}}^{(\tau -1)})=({\mathbf{u}}^{\left(\tau \right)},{\mathbf{v}}^{\left(\tau \right)},{\mathbf{p}}^{\left(\tau \right)},{\mathbf{q}}^{\left(\tau \right)})$ or $\tau =T_{max}$ 10:Output the suboptimal solution ${\mathbf{u}}^{*}={\mathbf{u}}^{\left(\tau \right)}$, ${\mathbf{v}}^{*}={\mathbf{v}}^{\left(\tau \right)}$, ${\mathbf{p}}^{*}={\mathbf{p}}^{\left(\tau \right)}$, ${\mathbf{q}}^{*}={\mathbf{q}}^{\left(\tau \right)}$