added the "initialize" method to Component

All instances of Component now have a method initialize.
This method must be called before any run of the optimization algorithm.

We now also test optimization algorithms more thoroughly and force them to have exactly equivalent runs for the same random seeds.
This actually required the introduction of the initialize method.

Added Jump Problem and Some Refacoring

Added Jump Problem and Some Refacoring

Added W-Model

Now, the W-Model benchmark problem is included.
The W-Model is a benchmark problem for discrete optimization with tunable ruggedness, deceptiveness, epistasis, and uniform neutrality.

minor bug fixes and improvements

minor bug fixes and improvements

Finalized Generalized EA and Its Components

The GeneralEA, its components, and the corresponding demo experiment are finished.

CMA-ES Algorithm Variants from Library `cmaes` Now Provided

We now also provide the CMA-ES algorithm variants from the library cmaes, which is developed by Masashi Shibata and Masahiro Nomura at https://github.com/CyberAgent/cmaes.
These algorithms are wrapped into the moptipy API and can now be accesses and experimented on in the same way as any other numerical optimization algorithm in our package.
This is shown in the small example continuous_optimization.py.

Fix of 0.9.23 to Comply with New `ruff` Rules

Fix of 0.9.23 to Comply with New ruff Rules

Bug Fixes and Added BOBYQA from PDFO

We now provide a wrapper around Powell's "Bound Optimization BY Quadratic Approximation" algorithm (BOBYQA) offered by the library "Powell's Derivative-Free Optimization solvers" (pdfo). This means that another highly efficient algorithm for numerical/continuous optimization is now available out of the box under our moptipy API.

We also included the first draft of an example for continuous optimization.

A set of strange bugs were fixed in StatRun and Ert.
There, we removed the numba jitting where it was not useful and problematic and fixed issues the accidental mismatch of Python ints and numpy ints.
We also better deal with the special case where StatRuns only have single values.

minor improvements and documentation fix

minor improvements and documentation fix

Removed bounds.py Utils: All Vector Spaces are now Bounded

All vector spaces are now bounded by finite box constraints.
We removed the bounds.py utility module and merged the functionality into the corresponding spaces.
The numerical optimization algorithms imported from SciPy have been updated accordingly.
This should lead to a more reasonable and maintainable API for numerical / continuous optimization.