get_sim = Model( RW, build_initial_model=True )
get_sim.collect_transition_data( num_det_calls=100, num_stoic_calls=10000 )
RW.layout.s_hash_print()
#get_sim.num_calls_layout_print()
#get_sim.min_num_calls_layout_print()
env = EnvBaseline( s_hash_rowL=RW.s_hash_rowL,
x_axis_label=RW.x_axis_label,
y_axis_label=RW.y_axis_label )
get_sim.add_all_data_to_an_environment( env )
policy, state_value = dp_value_iteration( env, do_summ_print=True, fmt_V='%.3f', fmt_R='%.1f',
max_iter=1000, err_delta=0.0001,
gamma=0.9, iteration_prints=10)
policy.save_diagram( RW, inp_colorD=None, save_name='dp_rw1000_policy',
show_arrows=False, scale=0.5, h_over_w=0.8,
show_terminal_labels=False)
print( 'Total Time =',time.time() - start_time )
pickle_esp.save_to_pickle_file( fname='dp_soln_to_randwalk_1000',
env=env, state_values=state_value, policy=policy)
show_start_policy=True,
max_iter=1000,
err_delta=0.0001,
gamma=0.9)
diag_colorD = {
'5': 'r',
'4': 'g',
'3': 'b',
'2': 'c',
'1': 'y',
'0': 'w',
'-5': 'r',
'-4': 'g',
'-3': 'b',
'-2': 'c',
'-1': 'y'
}
policy.save_diagram(env,
inp_colorD=diag_colorD,
save_name='car_rental_var_rtn',
show_arrows=False,
scale=0.25,
h_over_w=0.8)
pickle_esp.save_to_pickle_file(fname='dp_car_rental_PI',
env=env,
state_values=state_value,
policy=policy)
from introrl.utils import pickle_esp
RW = RandomWalk_1000Simulation()
model = Model(RW, build_initial_model=True)
model.collect_transition_data(num_det_calls=100, num_stoic_calls=10000)
print('Model Built')
# build an EnvBaseline from the Simulation
env = EnvBaseline(s_hash_rowL=RW.s_hash_rowL,
x_axis_label=RW.x_axis_label,
y_axis_label=RW.y_axis_label)
model.add_all_data_to_an_environment(env)
policy = Policy(environment=env)
policy.intialize_policy_to_equiprobable(env=env)
state_value = StateValues(env)
state_value.init_Vs_to_zero()
dp_policy_evaluation(policy,
state_value,
do_summ_print=True,
max_iter=1000,
err_delta=0.0001,
gamma=1.0)
pickle_esp.save_to_pickle_file(fname='random_walk_1000_PI_eval',
env=env,
state_values=state_value,
policy=policy)
env = EnvBaseline(s_hash_rowL=CR.s_hash_rowL,
x_axis_label=CR.x_axis_label,
y_axis_label=CR.y_axis_label)
get_sim.add_all_data_to_an_environment(env)
#env.save_to_pickle_file('car_rental')
#print('Saved env to *.env_pickle file')
print('built environment')
print('_' * 55)
#env.summ_print()
policy, state_value = dp_value_iteration(env,
do_summ_print=True,
fmt_V='%.1f',
fmt_R='%.1f',
max_iter=1000,
err_delta=0.0001,
gamma=0.9,
iteration_prints=10)
print('Total Time =', time.time() - start_time)
#env.save_to_pickle_file('car_rental')
pickle_esp.save_to_pickle_file(fname='car_rental_sim_to_env_const_rtn',
env=env,
state_values=state_value,
policy=policy)
get_sim.add_all_data_to_an_environment(env)
print('built environment')
print('_' * 55)
#env.summ_print()
policy, state_value = dp_value_iteration(env,
do_summ_print=True,
fmt_V='%.1f',
fmt_R='%.1f',
max_iter=1000,
err_delta=0.0001,
gamma=0.9,
iteration_prints=10)
policy.save_diagram(BJ,
inp_colorD=None,
save_name='dp_blackjack_policy',
show_arrows=False,
scale=0.5,
h_over_w=0.8,
show_terminal_labels=False)
print('Total Time =', time.time() - start_time)
pickle_esp.save_to_pickle_file(fname='dp_soln_to_blackjack',
env=env,
state_values=state_value,
policy=policy)