Polynomial Learning utility.

Polynomial Learning. More...

Files
file	polylearn.c
	Polynomial Learning - source file.
file	polylearn.h
	Polynomial Learning - header file.
Defines
#define	BUF_SIZE   512
	Buffer size.
#define	AWIN   1
	Value of an Alpha win.
#define	BWIN   5
	Value of a Beta win.
#define	DRAW   0
	Value of a draw.
#define	NUM_COEF   3
	Maximum value of a coefficient.
Functions
static void	_init (int argc, char *argv[])
	Perform initialization for Polylearn.
static void	_usage (void)
	Print usage message for Polylearn.
int	play (int alpha_first)
	Play a game between Alpha and Beta.
Variables
MOVE	o_move_1
	Move chosen by get_move(). Used by learn().
int	debug_1
	Is the debug flag on?
int	a_coeff [NUM_COEF]
	Alpha coefficients.
int	b_coeff [NUM_COEF]
	Beta coefficients.
static int	session
	Maximum of PLAYOFF_LEN sessions.
static int	n_flag
	Show games but do not play them?

---

Detailed Description

Polynomial Learning.

Apply the learning algorithm to PLAYOFF_LEN number of different sets of coefficients. Each coefficient set is viewed as a 3-tuple. PLAYOFF_LEN = (NUM_VALUES * NUM_VALUES * NUM_VALUES) and INTERVAL_SZ = (MAX_COEF / NUM_VALUES). Thus we can get a uniform distribution of 3-tuples by restricting coefficient values to be factors of INTERVAL_SZ.

usage: polylearn [-a] [-n] [-s] [-t 1|2]

-a : use aggression with polynomial. -n : show, but don't play, games. -s : starting session number. -t : apply trap heuristic (1 or 2).

---

Define Documentation

#define AWIN   1

Value of an Alpha win.

#define BUF_SIZE   512

Buffer size.

#define BWIN   5

Value of a Beta win.

#define DRAW   0

Value of a draw.

#define NUM_COEF   3

Maximum value of a coefficient.

---

Function Documentation

static void _init (  int  argc, char *  argv[] )  [static]

Perform initialization for Polylearn. Parameters: argc  int : sizeof argv[]. argv  char * [] : command line arguments. Returns: void.

static void _usage (  void   )  [static]

Print usage message for Polylearn. Returns: void.

play (  int  alpha_first  )

Play a game between Alpha and Beta. Parameters: alpha_first  int : flag - does Alpha go first? Returns: int. Return values: AWIN  - Alpha won the game. BWIN  - Beta won the game. DRAW  - cat game.

---

Variable Documentation

a_coeff

Alpha coefficients.

b_coeff

Beta coefficients.

debug_1

Is the debug flag on?

n_flag [static]

Show games but do not play them?

o_move_1

Move chosen by get_move(). Used by learn().

session [static]

Maximum of PLAYOFF_LEN sessions.

---