placement.py is a module containing code that will place routers in optimal
locations according to the metrics throughput and power. The key method in
this module is placement which takes a 2D matrix which describes device
population, the size of the room, the number of routers to be placed, which
metric to be optimized, whether the routers should be placed on the perimeter of
the room or anywhere on the grid, and placed routers which will not be moved
(this last argument was useful for another idea, but was ultimately abandoned. See
the note on placement_2 below. The
default argument is used for the results presented in the paper).
placement is run by optimal_placement which just runs placement
repeat-many times and takes the results which maximizes the metric.
optimal_placement is run by simulate which just uses optimal_placement
over different configurations (different number of routers, 2D vs. 1D placement,
different population scenarios, and the different metrics).
Any other functions below the functions described just supports the above functions in some capacity.
(simulate_2 and placement_2 were just different ideas to get an optimal
placement. It was found that placement worked better, so these functions
are really just vestiges now which don't affect anything, but are left in the
code to illustrate a different approach to the problem. simulate_2.py can be
ignored as well).
arrays.py contains the population scenarios used in the code. These are just
some ideas on possible population densities. These can be changed by just
changing what is fed into placement or one of the higher-level methods which
calls it, like simulate. This advice holds for introducing new metrics or
other ways of looking at the problem of router placement.
simulate.py is run as
simulate.py path repeat
where path is the file that the data from the simulations will be saved to and repeat is the number of times that a local maximum will be found for a
given simulated population (repeating more will probably yield a better result
as a global maximum is more likely to be found).
The output is a file which is a pickled array (using Python's pickle). The
structure of the array can be understood by just looking at the plot.py
plotting routine. It doesn't need to be understood in order to see results,
though. To see the results generated from running simulate.py, run
plot.py path
where path is the output file from simulate.py. You will get images named
power.png and throughput.png, samples of which are included.