def pomdp_choose_action( pomdp, Q, belief_for_policy, action_code_list ):
new_machine_action_number = tbh_pomdp_util.solve_qmdp( belief_for_policy, pomdp, Q )
new_machine_action = action_code_list[new_machine_action_number]
#print "ACTION NO., ACTION", new_machine_action_number, new_machine_action
# Determine if we need to reset the policy
reset_belief = determine_reset_belief( new_machine_action )
if reset_belief:
final_belief = tbh_pomdp_util.reset_belief( pomdp )
else:
final_belief = belief_for_policy
pomdp_dict = { 'belief_for_policy': belief_for_policy,
'final_belief': final_belief,
'new_machine_action': new_machine_action,
'reset_belief': reset_belief }
return pomdp_dict
def pomdp_choose_action( pomdp, Q, belief_for_policy, action_code_list ):
new_machine_action_number = tbh_pomdp_util.solve_qmdp( belief_for_policy, pomdp, Q )
new_machine_action = action_code_list[new_machine_action_number]
print "NEW MACHINE ACTION"
print new_machine_action
# Determine if we need to reset the policy
#reset_belief = determine_reset_belief( new_machine_action )
#if reset_belief:
# final_belief = tbh_pomdp_util.reset_belief( pomdp )
#else:
# final_belief = belief_for_policy
final_belief = belief_for_policy
reset_belief = False
pomdp_dict = { 'belief_for_policy': belief_for_policy,
'final_belief': final_belief,
'new_machine_action': new_machine_action,
'reset_belief': reset_belief }
return pomdp_dict
