Improvements to Function Minimization Example

[MashAliK]

Handful of changes to the Function Minimization example that improve robustness during evolution and result in a (in my opinion) better search function at the end. Summary of changes:

• During a run the program got stuck at iteration ~900 so I made a change to use multiprocessing instead of concurrent.features because the latter uses threads which, turns out, can't actually be killed from within the control loop
• Add a metric which accounts for the standard deviation of the x and y values. Without these, the values tend to fluctuate too much when calling the generated search functions
• Removed unnecessary calls to safe_float and float
• Removed value and distance from evaluate_stage1 since they aren't metrics (might not have values between 0 and 1)
• Updated the README to include an example generated with the new evaluator and included its results
• Exclude evaluate_function from Evolve Block scope because this function shouldn't be modified

[codelion]

@MashAliK thanks for the contribution can you fix the linting errors by running black openevolve tests examples as mentioned in https://github.com/codelion/openevolve/blob/main/CONTRIBUTING.md#code-style