Draft: Resolve "Implement multiple runs for one experiment"

Implements the last 2 points of the Issue:

• the new run should have a different scenario than the old one. You need to adapt ScenarioManager to randomly sample the bounds at the beginning of each new run. For this, look at config->scenario_setup. adapt the code so that trigger_interval: 0 only triggers at the beginning. You need to adapt the ScenarioManager class (_update_target method)
• provide an example file that adapts the standard example