readme

best routing protocol

Note: 

* "cr or PCR" was the internal name for S4 before the paper got
published. Don't be confused. :) 

* The current implementation was tested on Matlab 6 student version in
Linux, a rather old version of Matlab.  

* It will take a long time to get the simulation done. The following
optimization is used to speed up: (1) C implementation for the
dijkstra algorithm and BVR algorithm. Use make to compile it. (2) use
condor to run simulation instances in parallel. This is specifically
designed for UT Austin condor environment. Feel free to change the
scripts accordingly to fit your requirement, or run sequentially if
you have a really powerful machine and a good computer game to keep
you busy.

How to run the simulations:

1. varying number of beacons:

make sure high_density dir is there and empty, then in MATLAB

sim_parallel('high_density')

this will schedule a bunch of jobs in condor, use condor_q to check
the status of the jobs, and wait until they finish.

sim_parallel('high_density', 0) if you don't have condor and decide to
run it sequentially. It will take a looooooooooooong time to finish.

when the results are ready, run

sim_parallel('high_density', -1) to plot figures.

2. Directory setup traffic related (somehow badly named as DHT):

do not call sim_dht directly

sim_parallel('high_dht') 

sim_parallel('high_dht', -1)

3. varying network density

sim_parallel('low_density')
sim_parallel('low_density', -1) % to generate the all-in-one .mat file

sim_density will be called automatically

4. vary the size of the network
Note: sim_size is deprecated

sim_sizep('size')
sim_sizep('size', -1)

5. obstacles

sim_obstacle is the sequential version

sim_obstaclep('obstacle')
sim_obstaclep('obstacle',-1) is the parallel version


6. failure
sim_failure('low_rate')
sim_failure('low_rate', -1)

sim_failure('beacon_failure')
sim_failure('beacon_failure', -1)

7. incremental distance vector
sim_dv('high_dv')
sim_dv('high_dv', -1)