list_average_queuelength = []
list_average_queuingtimes = []
all_queue_lengths_overtime = np.zeros((len(diff_serverns), end_n_actions + 1))
queuelengthforrepetitios = np.zeros((n_simulations, end_n_actions + 1))
# run the simulation multiple times
i = 0
for n_server in diff_serverns:
queuelengthforrepetitios = np.zeros((n_simulations, end_n_actions + 1))
for j in range(0, n_simulations):
mu = 0.80
l = 0.64 * n_server
# initialize the global lists
init_global(end_n_actions)
# create a simpy environment
env = simpy.Environment()
# set up the system
env.process(
setup(env, n_server, mu, l, sjf, end_n_actions, db_helptime,
LT_value))
# run the program
env.run()
average_queuelength = np.average(global_variables.queue_length_list)
list_average_queuelength.append(average_queuelength)
n_batches = (end_n_actions-initialisation_period)/batch_size/2.
sjf = False # True to use shortest job first
'''Run simulation for different values of rho'''
list_nf_confidence_average_queuetimes = []
list_total_average_queuetimes = []
mu_range = np.arange(l+0.01, 0.8, 0.05)
for mu in mu_range:
list_average_queuelength = []
list_average_queuingtimes = []
# run the simulation multiple times
for i in range(n_simulations):
# initialize the global lists
init_global()
# create a simpy environment
env = simpy.Environment()
# set up the system
env.process(setup(env, n_server, mu, l, sjf, end_n_actions, "M", LT_value))
# run the program
env.run()
average_queuelength = np.average(global_variables.queue_length_list)
list_average_queuelength.append(average_queuelength)
list_batch_averages = batch_averages(batch_size, initialisation_period)
average_queuingtimes = np.average(global_variables.time_spend_in_queue_list)