I need to build an expression for the objective to be minimized which is based on the L1-norm. For doing this, I need to take the absolute value of a BQM, which is not possible, at least at the moment (switching to a L2-norm would be a good alternative, but squaring a BQM is taking very long time - my problem is large!).

I know the absolute value can be substituted with two constraints (https://stackoverflow.com/a/64109078/6498753), but how would you implement this in the context of QA? Would the equivalent of the dummy variable z be a new BQM to be added to the 'general' one (with something like BQM.update(Abs_Constraint1) and BQM.update(Abs_Constraint1))?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.