Replace normalization and auto_normalize keyes with scaling and auto_scale in ObjectiveFunctionConfig

[DanSava]

Issue
Resolves #9770

Approach
Short description of the approach

(Screenshot of new behavior in GUI if applicable)

• PR title captures the intent of the changes, and is fitting for release notes.
• Added appropriate release note label
• Commit history is consistent and clean, in line with the contribution guidelines.
• Make sure unit tests pass locally after every commit (git rebase -i main --exec 'just rapid-tests')

When applicable

• When there are user facing changes: Updated documentation
• New behavior or changes to existing untested code: Ensured that unit tests are added (See Ground Rules).
• Large PR: Prepare changes in small commits for more convenient review
• Bug fix: Add regression test for the bug
• Bug fix: Create Backport PR to latest release

[codspeed-hq]

CodSpeed Performance Report

Merging #9965 will not alter performance

_{Comparing DanSava:objectives-scaling (55a7074) with main (15f6077)}

Summary

✅ 25 untouched benchmarks

[verveerpj]

The idea is to change from a multiplication to a division factor. This is because we tend to have large objectives and then it is easier to specify the expected maximum value than its inverse. Division is also consistent with the scaling we use with constraints. So, in doc strings multiplication should be replaced with division, and in everest2ropt the 1.0 / objective.scaling should be replaced by objective.scaling

Otherwise looks good.

[verveerpj]

I think there also some test configuration yaml files that need changing?

[DanSava]

Division is also consistent with the scaling we use with constraints. So, in doc strings multiplication should be replaced with division, and in everest2ropt the 1.0 / objective.scaling should be replaced by objective.scaling

I am guessing currently ropt is not designed to handled ropt_config["objectives"]["scales"] = [ 1.0 / objective.scaling] because if I add this change then the output looks like this for config_multiobj.yml

 "single_objective_values": ['
   {'
      "accepted": 1,\n'
      "batch": 0,\n'
      "distance_p": -0.33333333,\n'
      "distance_q": -1.58333333,\n'
      "objective": -2.33333333\n'
    },
....
  ]

where previously it was

 "single_objective_values": ['
   {'
      "accepted": 1,
      "batch": 0,
      "distance_p": -0.75,
      "distance_q": -1.58333333,
      "objective": -2.33333333
    },
....
  ]

which seems a bit more consistent because -0.75 - 1.5833333 = -2.33333333 and -0.33333333 - 1.5833333 != -2.33333333

So I am a bit confused how both of these can be correct.

[DanSava]

Division is also consistent with the scaling we use with constraints. So, in doc strings multiplication should be replaced with division, and in everest2ropt the 1.0 / objective.scaling should be replaced by objective.scaling

I am guessing currently ropt is not designed to handled ropt_config["objectives"]["scales"] = [ 1.0 / objective.scaling] because if I add this change then the output looks like this for config_multiobj.yml

 "single_objective_values": ['
   {'
      "accepted": 1,\n'
      "batch": 0,\n'
      "distance_p": -0.33333333,\n'
      "distance_q": -1.58333333,\n'
      "objective": -2.33333333\n'
    },
....
  ]

where previously it was

 "single_objective_values": ['
   {'
      "accepted": 1,
      "batch": 0,
      "distance_p": -0.75,
      "distance_q": -1.58333333,
      "objective": -2.33333333
    },
....
  ]

which seems a bit more consistent because -0.75 - 1.5833333 = -2.33333333 and -0.33333333 - 1.5833333 != -2.33333333

So I am a bit confused how both of these can be correct.

Seems there was a place in the everest_data_api where the single objective values where calculated and I had also to update the formulate there to divide by the scaling value and not multiply with the normalization.

[verveerpj]

Division is also consistent with the scaling we use with constraints. So, in doc strings multiplication should be replaced with division, and in everest2ropt the 1.0 / objective.scaling should be replaced by objective.scaling

I am guessing currently ropt is not designed to handled ropt_config["objectives"]["scales"] = [ 1.0 / objective.scaling] because if I add this change then the output looks like this for config_multiobj.yml

It should be ropt_config["objectives"]["scales"] = [ objective.scaling] , so indeed no division there.

[verveerpj]

Seems there was a place in the everest_data_api where the single objective values where calculated and I had also to update the formulate there to divide by the scaling value and not multiply with the normalization.

That seems correct.

[verveerpj]

LGTM!