ROL: Access the updated x values for each iteration in the optimization process

[ofmla]

I would like to plot convergence curves, such as ||x_computed - x_true||/||x_true||, in addition to the objective/loss function curves. However, I couldn’t find an example of how to achieve this in ROL. How can I obtain this information when using ROL?Consider a simple 2D Rosenbrock problem. I want to use ROL to minimize the objective function and plot the convergence ||x_computed - x_true||/||x_true|| over iterations. Can someone provide an example code snippet or guidance on how to achieve this in ROL?

[ccober6]

@trilinos/rol

[dpkouri]

The objective function (ROL::Objective) has a function called "update" that you can use to output information. For an example of this, please see

packages/rol/example/PDE-OPT/topo-opt/src/filtered_compliance_robj.hpp