iterating over agents

[AtiyahElsheikh]

employing dox! simulation function (i.e. model based transition) obviously delivers less performance than the corresponding transition_based simulation function, i.e. x!. This hints that there is a better way to iterate over agents rather than the current default implementation given by all_people or alive_people,