Add: Initial implementation of polar_approx and unit tests

[mariam851]

Implemented a new hypervolume approximation algorithm using Polar Coordinates (Deng & Zhang approach) as requested in issue #582. This method improves sampling efficiency in high-dimensional spaces.

2. Technical Changes
   Core Implementation: Added polar_approx class in src/utils/hv_algos/.

Ray Generation: Implemented generate_random_ray() using std::normal_distribution to ensure uniform sampling on the unit hypersphere surface.

Normalization: Added logic to ensure all generated rays are Unit Vectors and stay within the positive orthant.

3. Build & Test
   CMake: Updated tests/CMakeLists.txt to register the new test case.

Unit Tests: Added a new test file to verify the mathematical properties of the generated rays (norm ≈ 1.0 and values ≥ 0).

[sylvesterkaczmarek]

Both the new implementation and test include pagmo/utils/hv_algos/hv_polar_approx.hpp, but that header is neither in master nor added by this PR, so the build fails before the new code can compile. Add the public header declaring polar_approx and keep a build/install test covering that include path.