def test_TIC(self):
"""
Test the TimeIndependentCounter
"""
tic = TimeIndependentCounter()
tic.count(3)
tic.count(2)
tic.count(5)
tic.count(0)
self.assertEqual(tic.get_mean(), 2.5,
msg="Error in TimeIndependentCounter. Wrong mean calculation or wrong counting.")
self.assertEqual(tic.get_var(), numpy.var([3, 2, 5, 0], ddof=1),
msg="Error in TimeIndependentCounter. Wrong variance calculation or wrong counting.")
self.assertEqual(tic.get_stddev(), numpy.std([3, 2, 5, 0], ddof=1),
msg="Error in TimeIndependentCounter. Wrong std dev calculation or wrong counting.")
tic.reset()
tic.count(3.)
tic.count(2.)
tic.count(5.)
tic.count(0.)
self.assertEqual(tic.get_mean(), 2.5,
msg="Error in TimeIndependentCounter. Wrong mean calculation or wrong counting.")
self.assertEqual(tic.get_var(), numpy.var([3, 2, 5, 0], ddof=1),
msg="Error in TimeIndependentCounter. Wrong variance calculation or wrong counting.")
self.assertEqual(tic.get_stddev(), numpy.std([3, 2, 5, 0], ddof=1),
msg="Error in TimeIndependentCounter. Wrong std dev calculation or wrong counting.")
def do_simulation_study(sim,
print_queue_length=False,
print_waiting_time=True):
"""
This simulation study is different from the one made in assignment 1. It is mainly used to gather and visualize
statistics for different buffer sizes S instead of finding a minimal number of spaces for a desired quality.
For every buffer size S (which ranges from 5 to 7), statistics are printed (depending on the input parameters).
Finally, after all runs, the results are plotted in order to visualize the differences and giving the ability
to compare them. The simulations are run first for 100s, then for 1000s. For each simulation time, two diagrams are
shown: one for the distribution of the mean waiting times and one for the average buffer usage
:param sim: the simulation object to do the simulation
:param print_queue_length: print the statistics for the queue length to the console
:param print_waiting_time: print the statistics for the waiting time to the console
"""
# counters for mean queue length and waiting time
counter_mean_queue_length = TimeIndependentCounter()
hist_mean_queue_length = TimeIndependentHistogram(sim, "q")
counter_mean_waiting_time = TimeIndependentCounter()
hist_mean_waiting_time = TimeIndependentHistogram(sim, "w")
# step through number of buffer spaces...
for S in sim.sim_param.S_VALUES:
sim.sim_param.S = S
counter_mean_queue_length.reset()
hist_mean_queue_length.reset()
counter_mean_waiting_time.reset()
hist_mean_waiting_time.reset()
sim.sim_param.SIM_TIME = 100000
sim.sim_param.NO_OF_RUNS = 1000
# repeat simulation
for run in range(sim.sim_param.NO_OF_RUNS):
# print(run)
sim.reset()
sim.do_simulation()
# add simulation result to counters and histograms (always use the mean)
counter_mean_queue_length.count(
sim.counter_collection.cnt_ql.get_mean())
hist_mean_queue_length.count(
sim.counter_collection.cnt_ql.get_mean())
counter_mean_waiting_time.count(
sim.counter_collection.cnt_wt.get_mean())
hist_mean_waiting_time.count(
sim.counter_collection.cnt_wt.get_mean())
pyplot.subplot(221)
pyplot.xlabel("Mean waiting time [ms] (SIM_TIME = 100.000ms)")
pyplot.ylabel("Distribution over n")
hist_mean_waiting_time.report()
pyplot.subplot(222)
pyplot.xlabel("Mean queue length (SIM_TIME = 100.000ms)")
pyplot.ylabel("Distribution over n")
hist_mean_queue_length.report()
# if desired, print statistics for queue length and waiting time
if print_queue_length:
print('Buffer size: ' + str(sim.sim_param.S) +
', simulation time: ' + str(sim.sim_param.SIM_TIME) +
', Mean buffer content: ' +
str(counter_mean_queue_length.get_mean()) + ' Variance: ' +
str(counter_mean_queue_length.get_var()))
if print_waiting_time:
print('Buffer size: ' + str(sim.sim_param.S) +
', simulation time: ' + str(sim.sim_param.SIM_TIME) +
', Mean waiting time: ' +
str(counter_mean_waiting_time.get_mean()) + ' Variance: ' +
str(counter_mean_waiting_time.get_var()))
counter_mean_queue_length.reset()
hist_mean_queue_length.reset()
counter_mean_waiting_time.reset()
hist_mean_waiting_time.reset()
sim.sim_param.SIM_TIME = 1000000
sim.sim_param.NO_OF_RUNS = 1000
# repeat simulation
for run in range(sim.sim_param.NO_OF_RUNS):
# print(run)
sim.reset()
sim.do_simulation()
# add simulation result to counters and histograms (always use the mean)
counter_mean_queue_length.count(
sim.counter_collection.cnt_ql.get_mean())
hist_mean_queue_length.count(
sim.counter_collection.cnt_ql.get_mean())
counter_mean_waiting_time.count(
sim.counter_collection.cnt_wt.get_mean())
hist_mean_waiting_time.count(
sim.counter_collection.cnt_wt.get_mean())
pyplot.subplot(223)
pyplot.xlabel("Mean waiting time [ms] (SIM_TIME = 1.000.000ms)")
pyplot.ylabel("Distribution over n")
hist_mean_waiting_time.report()
pyplot.subplot(224)
pyplot.xlabel("Mean queue length (SIM_TIME = 1.000.000ms)")
pyplot.ylabel("Distribution over n")
hist_mean_queue_length.report()
# if desired, print statistics for queue length and waiting time
if print_queue_length:
print('Buffer size: ' + str(sim.sim_param.S) +
', simulation time: ' + str(sim.sim_param.SIM_TIME) +
', Mean buffer content: ' +
str(counter_mean_queue_length.get_mean()) + ' Variance: ' +
str(counter_mean_queue_length.get_var()))
if print_waiting_time:
print('Buffer size: ' + str(sim.sim_param.S) +
', simulation time: ' + str(sim.sim_param.SIM_TIME) +
', Mean waiting time: ' +
str(counter_mean_waiting_time.get_mean()) + ' Variance: ' +
str(counter_mean_waiting_time.get_var()))
# set axis ranges for better comparison and display accumulated plot
pyplot.subplot(221)
pyplot.xlim([0, 3500])
pyplot.subplot(223)
pyplot.xlim([0, 3500])
pyplot.subplot(222)
pyplot.xlim([-.5, sim.sim_param.S_MAX + .5])
pyplot.subplot(224)
pyplot.xlim([-.5, sim.sim_param.S_MAX + .5])
pyplot.show()
