Evacuation processors return *always* positive time
They also have a positive value when:
- No agent reached the target
- Some agents are still in the scenario
- A source still spawns agents into the scenario
If a user is not aware of this property, then a wrong evacuation time (too small) is used. So it is extremely important that a the simulation time is set large enough. It is also not possible to detect whether all agents made it to the target.
When solving this issue also consider to add a check in
TestOptimizationMetricNelderMeadProcessor whether all agents reached the target.