claw::ai::game::select_random_action< Method > Class Template Reference

Select a random action among the best ones. More...

#include <game_ai.hpp>

List of all members.

Public Types
typedef Method::state	state
typedef Method::action	action
typedef Method::score	score
Public Member Functions
void	operator() (int depth, const state &current_state, action &new_action, bool computer_turn) const
	Select a random action among the best ones.

---

Detailed Description

template<typename Method>
class claw::ai::game::select_random_action< Method >

Select a random action among the best ones.

Template parameters:

• Method: the method to use to evaluate the actions.

Author:

Julien Jorge, Sébastien Angibaud

Definition at line 211 of file game_ai.hpp.

---

Member Typedef Documentation

template<typename Method >

typedef Method::action claw::ai::game::select_random_action< Method >::action

Definition at line 215 of file game_ai.hpp.

template<typename Method >

typedef Method::score claw::ai::game::select_random_action< Method >::score

Definition at line 216 of file game_ai.hpp.

template<typename Method >

typedef Method::state claw::ai::game::select_random_action< Method >::state

Definition at line 214 of file game_ai.hpp.

---

Member Function Documentation

template<typename Method >

void claw::ai::game::select_random_action< Method >::operator()	(	int	depth,
		const state &	current_state,
		action &	new_action,
		bool	computer_turn
	)		const

Select a random action among the best ones.

Parameters:

depth	Maximum depth of the search tree.
current_state	The state of the game.
new_action	(in/out) Best known action.
computer_turn	Tell if the action is done by the computer.

Definition at line 372 of file game_ai.tpp.

References claw::max_vector< E, Comp, Container >::add(), and claw::max_vector< E, Comp, Container >::get_v().

{
  std::list<action> l;
  typename std::list<action>::iterator it;
  action_eval<action, score> eval( new_action, current_state.min_score() );
  Method method;
  max_vector< action_eval<action, score> > events( eval );

  // get all reachable states
  current_state.next_actions( l );

  for (it=l.begin(); it!=l.end(); ++it)
    {
      state* new_state;
                        
      // try and evaluate each action
      new_state = static_cast<state*>(current_state.do_action(*it));

      eval.action = *it;
      eval.eval = method(depth-1, *new_state, !computer_turn);

      delete new_state;

      // keep the best actions.
      events.add( eval );
    }

  std::size_t i = (double)rand()/(RAND_MAX + 1) * events.get_v().size();
  new_action = events.get_v()[i].action;
} // select_random_action::operator()

---

The documentation for this class was generated from the following files:

• game_ai.hpp
• impl/game_ai.tpp