DiracDeltaDistribution

Models a Real Distribution backed by a Dirac Delta Function. This is similar to a Logistic probability function with a shape whose value tends to zero.

In practice, samples from this function return the provided value as a constant. The variance is zero, there is no randomness involved, and most of the useful information of a real distribution are actually lost. However, this utility can transform tools meant to work with a probability function in such a way that they work with a constant value (e.g., random walks with a constant step).