Load balancing framework missing about dgswemv2 HOT 2 OPEN

Currently the consensus, is that the strong typing required for HPX actions/ combined with the rather idiosyncratic behavior of the load balancers we will create will cause us to need to rely on strongly typed locality and world models. Specifically, we are looking at the following design choices:

1. a polymorphic representation of the submesh model; This allows us to break a circular dependency between locality and world models and the simulation units.
2. static locality and world models, which are defined on a per load balancer basis. Note special functions are required to ensure that they are properly instantiated and destroyed.

from dgswemv2.

Making implementation of semi-static load balancer a separate issue.

from dgswemv2.