Reduce default settings for solvers to minimum

[JesseJenkins]

We currently have fairly exhaustive sets of default settings for each solver copied into the configure_solver files for each solver. configure_highs.jl for example has 82 of them!

This raises the risk that the solvers change the keyword names of these settings and break GenX.

We should reduce this set of defaults to a minimum set of settings that are important to modify for GenX to run well on each solver. For example, whatever is required to use interior point/barrier method and to turn crossover off by default. We should also add a section to the documentation explaining these default configurations so users know if they want to change something they need to do so in the solver settings .yml files. Otherwise they can ignore solver settings .yml files.

Right now, we have some default settings in the example systems files for each solver settings .yml file. Some of these are correct and others probably are not. So rather than have these few key solver settings in the .yml examples, they should be moved to the solver settings .jl files and then the .yml files can be removed from the example folders, or used to show how one would change the default settings to something else (but commented out, so a user can see syntax example but it's not actually being used).

@sambuddhac happy to discuss this and decide in standard settings. This should be targeted to Develop branch I think (not Weekly Develop)

[sambuddhac]

Thanks @JesseJenkins . I will create a draft PR to address this and let's go over that on Monday, before finalizing.

[JesseJenkins]

@sambuddhac is this closed?

[sambuddhac]

@JesseJenkins , do think the list below is good enough? Right now, we still have the exhaustive list. But, I just took a closer look and it seems to me based on what we normally use and also considering the Benders decomposition work, this might suffice:

    "Feasib_Tol" => 1e-6,
    "Optimal_Tol" => 1e-4,
    "Pre_Solve" => "choose",
    "TimeLimit" => Inf,
    "Method" => "ipm",
    "parallel" => "choose",
    "ranging" => "off",
    "infinite_cost" => 1e+20,
    "infinite_bound" => 1e+20,
    "small_matrix_value" => 1e-09,
    "large_matrix_value" => 1e+15,
    "ipm_optimality_tolerance" => 1e-08,
    "objective_bound" => Inf,
    "objective_target" => -Inf,
    "random_seed" => 0,
    "threads" => 0,

Let me know what you think. Will draft a PR on this right now then.

[sambuddhac]

This issue will be addressed by PR #668

[sambuddhac]

This is now addressed on the release-0.4.0-rc branch