Improves the benchmark

[mbaudin47]

• Fills the README in order to better introduce the package.
• Moved the detailed analysis of each reliability problem into a dedicated "reliability_problems" directory.
• Performs a massive Monte-Carlo simulation for all the reliability problems in reliability_compute_reference_proba.ipynb
• Clarifies the benchmark of all methods on all reliability problems in reliability_benchmark_table.ipynb.
• Checks the consistency of the reference probabilities and the Monte-Carlo outputs in Check-reliability-reference-probabilities.ipynb.
• Computed the exact probability of RP55 up to approximately machine precision. This fixes #52
• Analysed the structure of RP75, tried FORM and explained why FORM may fail in RP75_Demo.ipynb
• Renamed the file OTReliabilityAlgorithmBenchmark to ReliabilityBenchmarkLibrary. Added doctstrings and unit tests.
• Added shapely requirement in order to compute the exact probability for RP55.
• Added tqdm requirement in order to compute the reference probabilities with massive Monte-Carlo.

[review-notebook-app[bot]]

Check out this pull request on 

Review Jupyter notebook visual diffs & provide feedback on notebooks.

---

Powered by ReviewNB

[mbaudin47]

• reliability_compute_reference_proba.ipynb: comment the export of the csv file so that overwriting the file by error cannot happen, add comments on the results.
• Check-reliability-reference-probabilities.ipynb: rerun the NB, add comments.
• reliability_benchmark_table.ipynb: add comments on the benchmark results

[mbaudin47]

The test passes, this is why I merge the PR.