raise ValueError
return ret
def pf_as_policy_type(i: int) -> Callable[[int], Sequence[str]]:
return get_sampling_func_from_prob_dict(policy_func(i))
this_qf = adp_pg_obj.get_act_value_func(pf_as_policy_type)
this_vf = adp_pg_obj.get_value_func(pf_as_policy_type)
print("Printing vf for a policy")
print(this_vf(1))
print(this_vf(2))
print(this_vf(3))
print("Printing DP vf for a policy")
from processes.policy import Policy
true_vf_for_pol = mdp_ref_obj1.get_value_func_dict(
Policy({s: policy_func(s)
for s in {1, 2, 3}}))
print(true_vf_for_pol)
opt_det_polf = adp_pg_obj.get_optimal_det_policy_func()
# noinspection PyShadowingNames
def opt_polf(s: S, opt_det_polf=opt_det_polf) -> Mapping[A, float]:
return {opt_det_polf(s): 1.0}
print("Printing Opt Policy")
print(opt_polf(1))
print(opt_polf(2))
print(opt_polf(3))
opt_vf = adp_pg_obj.get_value_func(adp_pg_obj.get_policy_as_policy_type())
# {s: policy_func(s) for s in {1, 2, 3}}
# ))
# print(true_vf_for_pol)
#
# # this_qf = adp_pg_obj.get_act_value_func_fa(policy_func)
# this_vf = adp_pg_obj.get_value_func_fa(policy_func)
# print("Printing vf for a policy")
# print(this_vf(1))
# print(this_vf(2))
# print(this_vf(3))
tol_val = 1e-6
true_opt = mdp_ref_obj1.get_optimal_policy(tol=tol_val)
print("Printing DP Opt Policy")
print(true_opt)
true_vf = mdp_ref_obj1.get_value_func_dict(true_opt)
print("Printing DP Opt VF")
print(true_vf)
opt_det_polf = adp_pg_obj.get_optimal_det_policy_func()
# noinspection PyShadowingNames
def opt_polf(s: S, opt_det_polf=opt_det_polf) -> Mapping[A, float]:
return {opt_det_polf(s): 1.0}
print("Printing Opt Policy")
print(opt_polf(1))
print(opt_polf(2))
print(opt_polf(3))
opt_vf = adp_pg_obj.get_value_func(adp_pg_obj.get_policy_as_policy_type())