Algorithm 2 EXPAND. Input: $\mathbf{\varphi }$ = $({\varphi }_1,\cdots ,{\varphi }_K)$, ${\Delta }_G$, $\Theta =[a_1,b_1]\times [a_2,b_2]\times ...\times [a_Q,b_Q]$, $\mathit{a}=[a_1,\cdots ,a_Q]$, $\mathit{b}=[b_1,\cdots ,b_Q]$, ${\Delta }_D$. Output: ${\mathbf{\varphi }}^{\prime }$=$({\varphi }_1^{\prime },\cdots ,{\varphi }_M^{\prime })$, where $M\le K(1+2Q)$. Note: In this algorithm, $\mathbf{\varphi }$ = $({\varphi }_1,\cdots ,{\varphi }_K)$ is a $Q\times K$ matrix, with $Q=dim\Theta$
	functionExpand($\mathbf{\varphi }$, ${\Delta }_G$, $\mathit{a}$, $\mathit{b}$, ${\Delta }_D$)
2:	Initialize: $[Q,K]=size\left(\mathbf{\varphi }\right)$, $\mathbf{I}=\mathbf{Q}\times \mathbf{Q}$ Identity matrix, $\mathrm{new}\mathbf{\varphi }⟵\mathbf{\varphi }$
	for $k=1,\dots ,K$ do	▹$K=\mathrm{number}\hspace{0.17em}\mathrm{of}\hspace{0.17em}\mathrm{of}\hspace{0.17em}\mathrm{input}\hspace{0.17em}\mathrm{support}\hspace{0.17em}\mathrm{points}$
4:	for $d=1,\dots ,Q$ do	▹$Q=dim\Theta$
	$T\left(d\right)={\Delta }_G(b\left(d\right)-a\left(d\right))$
6:	if $\varphi (d,k)+T\left(d\right)\le b\left(d\right)$ then	▹ Check upper boundary
	${\mathbf{\varphi }}^{+}=\mathbf{\varphi }(:,k)+T\left(d\right)\mathbf{I}(:,d)$
8:	$dist={10}^{30}$
	end if
10:	for $k_{in}=1:\mathrm{length}\left(\mathrm{new}\mathbf{\varphi }\right)$ do
	$\mathrm{newdist}=\sum \mathrm{abs}({\mathbf{\varphi }}^{+}-\mathrm{new}\mathbf{\varphi }(:,k_{in}))./(\mathit{b}-\mathit{a})$	▹ x ./y done
	component-wise
12:	$dist=min(dist,\mathrm{newdist})$
	end for
14:	if $dist\ge {\Delta }_D$ then	▹ Check distance to new support point
	$\mathrm{new}\mathbf{\varphi }⟵\left[\mathrm{new}\mathbf{\varphi },{\mathbf{\varphi }}^{+}\right]$
16:	end if
	if $\varphi (d,k)-T\left(d\right)\ge a\left(d\right)$ then	▹ Check lower boundary
18:	${\mathbf{\varphi }}^{-}=\mathbf{\varphi }(:,k)-T\left(d\right)\mathbf{I}(:,d)$
	$dist={10}^{30}$
20:	end if
	for $k_{in}=1:\mathrm{length}\left(\mathrm{new}\mathbf{\varphi }(1,:)\right)$ do
22:	$\mathrm{newdist}=\sum (abs({\mathbf{\varphi }}^{-}-\mathrm{new}\mathbf{\varphi }(:,k_{in}))./(b-a))$	▹ x./y done
	component-wise
	$dist=min(dist,\mathrm{newdist})$
24:	end for
	if $dist\ge {\Delta }_D$ then	▹ Check distance to new support point
26:	$\mathrm{new}\mathbf{\varphi }⟵[\mathrm{new}\mathbf{\varphi },{\mathbf{\varphi }}^{-}]$
	end if
28:	end for
	end for
30:	$\mathbf{\varphi }⟵new\mathbf{\varphi }$
	end function