Douglas-Rachford Spliting with an offset for monotone operators

22 Views Asked by At

The zero finding problem that the Douglas-Rachford Splitting (DRS) Algorithm attempts to solve has the following formulation

$0 \in (A+B)(x)$

where $A$ and $B$ are maximal monotone operators. Then, the DRS algorithm finds the corresponding zero of the equation above by splitting the computations using resolvents $J_{\alpha A}$ and $J_{\alpha B}$ as:

$ x^{k + \frac{1}{2}} = J_{\alpha B}(z^K)$

$ x^{k + 1} = J_{\alpha A}(2*x^{k + \frac{1}{2}} - z^k)$

$ z^{k + 1} = z^{k} + x^{k + 1} - x^{k + \frac{1}{2}}$,

which is clear. Now, my problem is how do we extend this algorithm if the zero finding problem has a constant offset so the problem can be written as

$cte \in (A+B)(x)$

where $cte$ is a constant value.