The contents of this report describes an implementation of the algorithm described in the still unpublished paper, Approximation algorithm for geometric shortest paths in weighted 3d-regions [Aleksandrov et al., 2010]. Details about how the Computational Geometry Algorithms Library (CGAL), Boost and Qt libraries were integrated into final solution are also given.

This report then provides an extensive overview of all the main challenges and steps that were undertaken in implementing the algorithm. This overview should provide the reader with a good idea of how a lot of the implementation was created.

An attempt to verify the correctness of the implementation is then made. This is done by ensuring that the theoretical upper bounds on the number of points that the algorithm should produce is higher than what the implementation actually creates. Other aspects of the implementation are also briefly verified.

This report displays the results of multiple performance tests on the implementation and their meaning. It is shown that various parameters that the implementation uses can be manipulated to allow average systems to use the algorithm practically. It is also shown that it is possible to run the algorithm with very high accuracy if a high end enough system is available.

Lastly, numerous possible extensions to the implementation that could provide significant performance boosts are outlined. These suggestions would be a very good starting point for any future work to be done.