PBDMST Algorithm.

Star when solved, newest on top
- ?RNG?: will multiplying RNG seed value by rank give sufficient distribution
  to the different processes?
* MPI_Gather() is definitely getting values correctly; just tried adding
  the following code in front of it and the correct values were passed:
      if(mpiRank == 1)
          bestCost = 10000;
      if(mpiRank == 2)
          bestCost = 20000;
      et cetera for processes 1 to n
- implement parallel optimization; right now, all 4 processes are doing
  optimization independantly
* there still appears to be a problem with how the information is being
  passed: processes 1 to n are returning identical bestCost values and
  node 0 is improving; is this an issue with the RNG?       
- MPI_Gather() wasn't working because the receive count is supposed to
  be set s.t. sendcount == recvcount, which isn't intuitive, but it's
  no longer hanging
- warning in processFile.h: 58, 84: eCount set but not used
* search ?02? in pbdmst.cxx: question about how edges are being
  initialized
* search ?01? in pbdmst.cxx: can NFS support concurrent reads for
  initializing the graphs? currently implemented so that each process
  steps through individually
* pass the main command line arguments; MPI requires argc and argv in
  MPI_Init()
* change main loop to run to full duration every time so we don't have
  to worry about processes finishing out of sync
* packPheromones() as it stands will write edges twice, one for each side
  of the edge