Use Poisson distribution for spawn generation and measure the real parameter value with data processor

Until now, we use the mean inter arrival time as parameter in the negative exponential distribution for agent generation.

New features

1. Use the Poisson distribution as an alternative to the negative exponential distribution, since both are related.
2. Measure the number of generated agents using the new implemented data processor NumberOfGeneratedPedsProcessor.

Note to point 2.:

If the number of generated agents is high, the mean of the NumberOfGeneratedPedsProcessor values should converge to the value of the Poisson distribution parameter of the source (provided that there is only one source in the scenario).