ray simulations seem to work

nodes (except head node) can deconnect and reconnect to/from the cluster and work is rebalanced properly. if a node dies while simulating, the leftover work is picked up at the end and everything is merged normall. Tests of the actual validity of the mergd data have not been made yet

README.md

@@ -185,8 +185,8 @@ tolerated_error: float
 step_size: float
     if given, sets a constant step size rather than adapting it.
 
-parallel: int
-    how many parallel simulations to run. default : 1
+parallel: bool
+    whether to run simulations in parallel with the available ressources. default : false
 
 repeat: int
     how many simulations to run per parameter set. default : 1