The failure probability of RP55 is wrong
According to:
https://rprepo.readthedocs.io/en/latest/reliability_problems.html#id76
the failure probability of RP55 is 0.36.
We checked this with various methods:
- Monte-Carlo: [5.60e-01,5.60e-01] at 95% (with 1.523630 * 10^9 points)
- Quasi-Monte-Carlo: 0.559386 (with 10^6 points)
- Gauss-Legendre quadrature: 0.558822 (with 10^4 points)
- FORM: 0.4992
Not all our methods were satisfactory on this easy problem, but the exact probability cannot be 0.36.