Add an example on how `probability_scale_factor` works, perhaps illustrating the benefit of it in the performance in the [n-monitoring](https://gamma-opt.github.io/DecisionProgramming.jl/dev/examples/n-monitoring/) example?