departitioning

[eagmon]

ecoli_master currently uses the partitioning assumption for its PartitionedProcesses, with the Allocator process. This has been compared to training-wheels, where were inherited from the wcEcoli model. Removing this assumption will allow the currently-partitioned processes to run independently at their own timescale.

Here are some proposed steps for departitioning:

1. Make EcoliMaster configurable to remove processes from the partitioned_processes list.
2. Remove one process at a time from the list, to identify the minimal set of processes tied together by partitioning.
3. Fix the remaining processes one at a time and remove them from partitioning.
4. Model the actual biology that results in the desired partitioned behavior. Possibly use a diffusion-limited processes to act as a mechanistic stand-in for allocation. Some parameters might also need to be improved.