service_rate = 1.0
bandwidth = 0.6
# Figure Plot Parameter
ecdf_samples = 10000
x_lim = 20.0
y_lim = 1.0
xy_lim = (x_lim, y_lim)
x_axis = np.linspace(0, x_lim, ecdf_samples)
# Generate Theoretical Curve
y_curves = []
y_theo = []
for item in x_axis:
value = MD1_response_CDF(arrival_rate, service_rate, item)
y_theo.append(value)
y_curves.append(y_theo)
#=============== Simulation ================
# Generate Emprical Samples
arrival_evt = gen_poisson_process(arrival_rate, sample_num)
# Stimulate the server
for period in [0.5, 1.0, 5.0, 10, 50]:
budget = period * bandwidth
(atserver_evt, leave_evt) = run_D_FIFO_DS_server(budget, period,
service_rate, arrival_evt)
response_time = np.subtract(leave_evt, arrival_evt)
response_time = np.subtract(leave_evt, arrival_evt)
ecdf1 = sm.distributions.ECDF(response_time)
y_empr_ds = ecdf1(x_axis)
#=============== For DS Worst ===============
# Stimulate the server
(atserver_evt, leave_evt) = run_D_FIFO_PS_server(budget, period, service_rate,
arrival_evt)
response_time = np.subtract(leave_evt, arrival_evt)
ecdf2 = sm.distributions.ECDF(response_time)
y_empr_ps = ecdf2(x_axis)
# Generate Theoretical Curve
y_theo = []
for item in x_axis:
value = MD1_response_CDF(arrival_rate, service_rate, item)
y_theo.append(value)
# Generate Theoretical Curve M/D/1 nerf
y_theo_l = []
for item in x_axis:
value = MD1_response_CDF(arrival_rate, service_rate * bandwidth, item)
y_theo_l.append(value)
#mytitle = u'P=2.0, Bw=65%, \u03BB=0.5, \u03BC=1.0'
mytitle = "P=%.1f, Bw=%2d%%, %c=%.1f, %c=%.1f" % (period, int(
bandwidth * 100), u'\u03BB', arrival_rate, u'\u03BC', service_rate)
#plot_curves_with_same_x(x_axis, [y_theo], ['M/D/1 Theoretical'], xy_lim, mytitle)
#plot_curves_with_same_x(x_axis, [y_theo, y_empr_ds], ['M/D/1 Theoretical', 'M/D(DS)/1 Best'], xy_lim, mytitle)
plot_curves_with_same_x(