For the beamline alignment problem, human experts typically optimize upstream components first and move down the beamline until they reach the endstation. Blop could write a Bluesky plan to incorporate something like this where a user can specify a set of optimization experiments to run, each describing the DOFs and Objectives for each section of the beamline. Within each section, it would optimize toward a local-minima, then once satisfied, it can move onto the next section.

• One experiment configuration per section
• Proceed to the next section based on some outcome threshold or by user approval
• Final global optimization step for fine-tuning the result