fo

.in
rs
de
ea
yr
www.myreaders.info/ , RC Chakraborty, e-mail [email protected] , June 01, 2010
www.myreaders.info/html/artificial_intelligence.html
www.myreaders.info

ha

kr

ab

or

ty

,w

.m

Game Playing : AI Course Lecture 29 30, notes, slides

Return to Website

Game Playing
Artificial Intelligence
Game

Playing,

topics

theory, relevance of

Overview,

definition

of

game,

game

game theory and game plying, Glossary of

terms game, player, strategy, zero-Sum game, constant-sum

game, nonzero-sum game, Prisoner's dilemma, N-Person game,
utility

function,

mixed

theorem, saddle point;

strategies,

expected

payoff,

Mini-Max

taxonomy of games. Mini-Max search

procedure : formalizing game - general and a Tic-Tac-Toe game,

evaluation function ;

MINI-MAX technique : game trees, Mini-Max

algorithm. Game Playing with Mini-Max : example of Tic-Tac-Toe moves,

obtained.

static evaluation,

back-up the evaluations and evaluation

Alpha-Beta Pruning : Alpha-cutoff, Beta-cutoff.

fo
.in
rs
de
ea
,w

.m

yr

Game Playing

ha

kr

ab

or

ty

Artificial Intelligence

Topics
(Lectures 29, 30,

Slides

2 hours)

1. Overview

03-18

Definition of Game, Game theory, Relevance of Game theory and Game

plying, Glossary of terms Game, Player, Strategy, Zero-Sum game,
Constant-Sum game, Nonzero-Sum game, Prisoner's dilemma, N-Person
Game, Utility function, Mixed strategies, Expected payoff, Mini-Max
theorem, Saddle point; Taxonomy of games.
2. Mini-Max Search Procedure

Formalizing game :

General

19-25

and

Tic-Tac-Toe

game, Evaluation

function ; MINI-MAX Technique : Game Trees, Mini-Max algorithm.

3. Game Playing with Mini-Max

26-32

Example : Tic-Tac-Toe - Moves, Static evaluation, Back-up the

evaluations, Evaluation obtained.
4. Alpha-Beta Pruning

33-35

Alpha-cutoff, Beta-cutoff
5. References
02

36

fo
.in
rs
de
ea
yr

What is Game ?

ha

kr

ab

or

ty

,w

.m

Game Playing

The term Game means a sort of conflict in which n individuals or

groups (known as players) participate.

Game theory denotes games of strategy.

John von Neumann is acknowledged as father of game theory. Neumann
defined Game theory in 1928 and 1937 and established the mathematical
framework for all subsequent theoretical developments.

Game theory allows decision-makers (players) to cope with other

decision-makers (players) who have different purposes in mind. In
other

words,

players

determine

their

own

strategies

in

terms

the strategies and goals of their opponent.

Games are integral attribute of human beings.

Games engage the intellectual faculties of humans.

If computers are to mimic people they should be able to play games.

03

of

fo
.in
rs
de
ea

Overview

or

ty

,w

.m

yr

1. Over View
Game playing, besides the topic of attraction

to

the people, has close

ha

kr

ab

relation to "intelligence", and its well-defined states and rules.

The most commonly used AI technique in game is "Search".

A "Two-person zero-sum game" is most studied game
have exactly opposite goals. Besides
(such as chess and Go)

and

where the two players

there are "Perfect information games"

"Imperfect information games" (such as bridge

and games where a dice is used).

Given sufficient time and space, usually an optimum solution can be obtained
for the former by exhaustive search, though not for the latter.
many

However, for

interesting games, such a solution is usually too inefficient to be

practically used.
Applications of game theory are wide-ranging. Von Neumann and Morgenstern
indicated the utility of game theory by linking with economic behavior.
Economic models : For markets of various commodities with differing
numbers of buyers and sellers, fluctuating values of supply and demand,
seasonal and cyclical variations, analysis of conflicts of interest in maximizing
profits and promoting the widest distribution of goods and services.
Social sciences : The n-person game theory has interesting uses in studying
the distribution of power in legislative procedures, problems of majority rule,
individual and group decision making.
Epidemiologists :

Make use of game theory, with respect to immunization

procedures and methods of testing a vaccine or other medication.

Military strategists : Turn to game theory to study conflicts of interest
resolved through "battles" where the outcome or payoff of a war game is
either victory or defeat.
04

fo
.in
rs
de
ea

Overview

or

ty

,w

.m

yr

1.1 Definition of Game

A game has at least two players.

kr

ab

Solitaire is not considered a game by game theory.

ha

The term 'solitaire' is used for single-player games of concentration.

An instance of a game begins with a player choosing from a set of specified

(game rules) alternatives. This choice is called a move.

After first move, the new situation determines which player to make next
move and alternatives available to that player.
In many board games, the next move is by other player.
In many multi-player card games, the player making next move
depends on who dealt, who took last trick, won last hand, etc.

The moves made by a player may or may not be known to other players.
Games in which all moves of all players are known to everyone are called
games of perfect information.
Most board games are games of perfect information.
Most card games are not games of perfect information.

Every instance of the game must end.

When an instance of a game ends, each player receives a payoff.
A payoff is a value associated with each player's final situation.
A zero-sum game is one in which elements of payoff matrix sum to zero.
In a typical zero-sum game :
win

draw =
loss
05

1 point,
0 points, and

= -1 points.

fo
.in
rs
de
ea

Overview

or

ty

,w

.m

yr

1.2 Game Theory

Game theory does not prescribe a way or say how to play a game.

kr

ab

Game theory is a set of ideas and techniques for analyzing conflict

ha

situations between two or more parties. The outcomes are determined

by their decisions.
General game theorem : In every two player, zero sum, non-random,
perfect knowledge game, there exists a perfect strategy guaranteed to
at least result in a tie game.
The frequently used terms :
The term "game" means a sort of conflict in which n individuals or
groups (known as players) participate.
A list of "rules" stipulates the conditions under which the game begins.
A game is said to have "perfect information" if all moves are known to
each of the players involved.
A "strategy" is a list of the optimal choices for each player at every
stage of a given game.
A "move" is the way in which game progresses from one stage to
another, beginning with an initial state of the game to the final state.
The total number of moves constitute the entirety of the game.
The payoff or outcome, refers to what happens at the end of a game.
Minimax - The least good of all good outcomes.
Maximin - The least bad of all bad outcomes.
The primary game theory is the Mini-Max Theorem. This theorem says :
"If a Minimax of one player corresponds to a Maximin of the other
player, then that outcome is the best both players can hope for."

06

fo
.in
rs
de
ea

Overview

or

ty

,w

.m

yr

1.3 Relevance Game Theory and Game Plying

How

relevant

the Game theory is

to Mathematics, Computer science

ha

kr

ab

and Economics is shown in the Fig below.

Game Playing
Games can be Deterministic or non-deterministic.
Games can have perfect information or imperfect information.

07

Games

Deterministic

Non- Deterministic

Perfect
information

Chess, Checkers, Go,

Othello, Tic-tac-toe

Backgammon,
Monopoly

Imperfect
information

Navigating
a maze

Bridge, Poker,
Scrabble

fo
.in
rs
de
ea

Overview

or

ty

,w

.m

yr

1.4 Glossary of terms in the context of Game Theory

Game

kr

ab

Denotes

games

of

strategy.

It

allows

decision-makers

(players)

ha

to cope with other decision-makers (players) who have different

purposes in mind. In other words, players determine their own

strategies in terms of the strategies and goals of their opponent.
Player
Could be one person, two persons or a group of people who share
identical interests with respect to the game.
Strategy
A player's strategy in a game is a complete plan of action for
whatever situation might arise. It is the complete description of
how one will behave under every possible circumstance. You need
to analyze the game mathematically and

create a table with

"outcomes" listed for each strategy.

A two player strategy table

08

Players
Strategies

Player A
Strategy 1

Player A
Strategy 2

Player A
Strategy 3

etc

Player B
Strategy 1

Tie

A wins

B wins

...

Player B
Strategy 2

B wins

Tie

A wins

...

Player B
Strategy 3

A wins

B wins

Tie

...

etc

...

...

...

...

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty

Zero-Sum Game

Overview

It is the game where the interests of the players are diametrically

ab

or

opposed. Regardless of the outcome of the game, the winnings of the

ha

kr

player(s) are exactly balanced by the losses of the other(s).

No wealth is created or destroyed.

There are two types of zero-sum games:
Perfect information zero-sum games
General zero-sum games
The difference is the amount of information available to the players.
Perfect Information Games :

Here all moves of all players are known to everyone.

e.g., Chess and Go;
General Zero-Sum Games :

Players must choose their strategies simultaneously, neither knowing

what the other player is going to do.
e.g., If you play a single game of chess with someone, one person
will lose and one person will win. The win (+1) added to the loss (-1)
equals zero.
09

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty

Overview

Constant-Sum Game

Here the algebraic sum of the outcomes are always constant,

ab

or

though not necessarily zero.

ha

kr

It is strategically equivalent to zero-sum games.

Nonzero-Sum Game
Here the algebraic sum of the outcomes are not constant. In these
games, the sum of the pay offs are not the same for all outcomes.
They are not always completely solvable but provide insights into
important areas of inter-dependent choice.
In these games, one player's losses do not always equal another
player's gains.
The nonzero-sum games are of two types:
Negative Sum Games (Competitive) : Here nobody really wins, rather,

everybody loses. Example - a war or a strike.

Positive Sum Games (Cooperative) :

Here all players have one goal

that they contribute together. Example - an educational game, building

blocks, or a science exhibit.
10

fo
.in
rs
de
ea
yr
or

ty

,w

.m

Overview

Prisoner's Dilemma
It is a two-person nonzero-sum game. It is a non cooperative game

kr

ab

because the players can not communicate their intentions.

ha

Example : The two players are partners in a crime who have been captured by

the police. Each suspect is placed in a separate cell and offered the
opportunity to confess to the crime.
Now set up the payoff matrix. The entries in the matrix are two numbers
representing the payoff to the first and second player respectively.
Players

2nd player
Not Confess

2nd player
Confess

1st player
Not Confess

5 ,5

0 , 10

1st player
Confess

10 , 0

1 , 1

If neither suspect confesses, they go free, and split the proceeds of their
crime, represented by 5 units of payoff for each suspect.

If one prisoner confesses and the other does not, the prisoner who
confesses testifies against the other in exchange for going free and gets the
entire 10 units of payoff, while the prisoner who did not confess goes to
prison and gets nothing.

If both prisoners confess, then both are convicted but given a reduced
term, represented by 1 unit of payoff : it is better than having just the
other prisoner confess, but not so good as going free.
This game represents many important aspects of game theory.
No matter what a suspect believes his partner is going to do, it is always best
to confess.

If the partner in the other cell is not confessing, it is possible to get 10

instead of 5 as a payoff.

If the partner in the other cell is confessing, it is possible to get 1 instead

of 0 as a payoff.

Thus the pursuit of individually sensible behavior results in each player

getting only 1 as a payoff, much less than the 5 which they would get if
neither confessed.

This conflict between the pursuit of individual goals and the common
good is at the heart of many game theoretic problems.
11

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty

N-Person Game

Overview

Involve more than two players.

ab

or

Analysis of such games is more complex than zero-sum games.

ha

kr

Conflicts of interest are less obvious.

Here, what is good for player-1 may be bad for player-2 but good
for player-3. In such a situation coalitions may form and change a
game radically. The obvious questions are :
how would the coalition form ?
who will form coalitions ?
would the weak gang up against the strong ? or

would weak players make an alliance with a strong ?

In either case it could be looked at as two coalitions, and so effectively
a two person game.
Utility Function
It is quantification of person's preferences respect to certain objects. In
any game, utility represents the motivations of players. A utility function
for a given player assigns a number for every possible outcome of the
game with the property that a higher number implies that the outcome
is more preferred.
12

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty

Overview

Mixed Strategies

A player's strategy in a game is a complete plan of action for

kr

ab

or

whatever situation might arise. It is a complete algorithm for playing

ha

the game, telling a player what to do for every possible situation

throughout the game.

A Pure strategy provides a complete definition of how a player will
play a game. In particular, it determines the move a player will make
for any situation they could face. A player's strategy set is the set of
pure strategies available to that player.
A Mixed strategy

is an assignment of a probability to each pure

strategy. This allows for a player to randomly select a pure strategy.

Since probabilities are continuous, there are infinitely many mixed
strategies available to a player, even if their strategy set is finite.
A mixed strategy for a player is a probability distribution on the set
of his pure strategies.
Suppose a player has only a finite number, m, of pure strategies,
then a mixed strategy reduces to an

m-vector,

X = (x1, , xm) ,

satisfying
xi 0 ,

i=1

xi

=1

Now denote the set of all mixed strategies for player-1 by X, and
the set of all mixed strategies for player-2 by Y.
X = { x = (xi , . . . , xm) : xi 0 ,

xi

=1

i=1

Y = { y = (yi , . . . , yn) : yi
[continued in the next slide]
13

0 ,

i=1

yi = 1

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty

Overview

[continuing from previous slide- mixed strategy]

Expected Payoff

ab

or

Suppose that player1 and player2 are playing the matrix game A.

ha

kr

If player1 chooses the mixed strategy xi, and

player2 chooses the mixed strategy yj,

then the expected payoff aij will be computed by
y1

a11
.
.
am1

X1

xm

.
.
.
.
.

.
.
.
.
.

That is

.
.
.
.
.

.
.
.
.
.

yn

y1

a1n
.
.
a1n

A(x , y) =

X1

X1a11 y1

Xm

Xmam1 y1

i=1

or in matrix form

.
.

j=1
T

.
.
.
.
.

..
..
..
..
...

.
.
.
.
.

yn
X1a1n yn

.
.

Xmamn yn

xi aij yj

A(x , y) = x A y

This can be thought as a weighted average of the expected payoffs.

Player1s

Maximin Strategy :

Assume that player1 uses x, and player2 chooses y to minimize A(x, y);
player1's expected gain will be
V (x) = min x
j

Aj

where Aj is jth column of the matrix A

Player1 should choose x so as to maximize V(x):

V1

= max min x
xX

Aj

Such a strategy x is player1s maximin strategy.

Player2s

Player2's

Minimax Strategy :

expected loss ceiling will be

T
V (y) = max Ai y
i

where Ai is ith row of the matrix A

Player2 should choose y so as to minimize V(y) :

T
V 2 = min max Ai y
yY

Such a strategy y is player2s

minimax strategy.

Thus we obtain the two numbers V1 and V2.

These numbers are

called the values of the game to player1 and player2, respectively.

14

fo
.in
rs
de
ea
yr
or

ty

,w

.m

Overview

Mini-Max Theorem [Ref. previous slide]

It is a concept of Games theory.

kr

ab

Players adopt those strategies which will maximize their gains, while

ha

minimizing their losses. Therefore the solution is, the best each

player can do for him/herself in the face of opposition of the other

player.
In the previous slide, the two numbers V1 and V2 are the values
of the game to player1 and player2, respectively.
Mini-Max Theorem :

V1 = V2

The minimax theorem states that for every two-person, zero-sum

game, there always exists a mixed strategy for each player such that
the expected payoff for one player is same as the expected cost
for the other.
This theorem is most important in game theory. It says that every
two-person zero-sum game will have optimal strategies.
15

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty

Overview

Saddle Point
An element

aij

of a

matrix

is

described as a saddle point

if

it

ha

kr

ab

or

equals both the minimum of row i and the maximum of column j.

A Saddle Point is a payoff that is simultaneously a row minimum

and a column maximum. In a game matrix, if the element aij
corresponds to a saddle point then it is the largest in its column
and the smallest in its row (LCSR).
The value aij is called the optimal payoff.
Examples : Three game payoff matrixes
A

5
3
-3

B
1
2
0

3
4
1

one saddle point

a22 = 2

4
0
6

3
1
3

5
0
9

two saddle points

a12 = 3 , a32 = 3

-1
1

1
-1

no saddle point

The game payoff matrix B shows that saddle point may not be
unique, but the optimal payoff is unique.
A payoff matrix needn't have a saddle point, but if it does, then
the usually minimax theorem is easily shown to hold true.
The Zero Sum Games modeled in matrices are solved by finding the
saddle point solution.
[continued in the next slide]
16

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty
or
ab

How to find Saddle Point ?

Below shown two examples : A and B

j

ha

kr

5
3
-3

1
2
0

3
4
1

Max

Overview

[continuing from the previous slide Saddle point]

min
1

2
-3

min

4
0
6

3
1
3

5
0
9

Max

one saddle point

a22 = 2

3
0

two saddle points

1. Find your min payoff : Label each row at its end with its minimum
payoff. This way you'll define your worst case scenarios when
choosing a strategy.
2. Find your opponent's min payoff : Label each column at its bottom
with its maximum payoff. This will show the worst case scenarios for
your opponent.
3. Find out which is the highest value in the series of minimum values.
It is at one place as 2 in matrix A and at two places as 3 in matrix B.
4. Then find out which is the lowest value in the series of maximum
values. It is as 2 in matrix A and as 3 in matrix B.
5. Find out if there is a minimax solution : If these two values match,
then you have found the saddle point cell.
6. If there is a minimax solution, then it is possible that both agents
choose the corresponding strategies. You and your opponent are
maximizing the gain that the worst possible scenario can drive.
17

fo
.in
rs
de
ea

Overview

All that are explained in previous section are summarized below.

kr

ab

or

ty

,w

.m

yr

1.5 Taxonomy of Games

ha

Game Theory
Games of Skill

Games of Chance

Games of Strategy

Games involving risk

Games involving uncertainty

Mixed-motive

Cooperative

Purely
cooperative

Minimal
social
situation

Perfect info

Infinite

Two-persons

Zero-sum

Coalitions not
permitted

Finite

Noncooperative

Cooperative

Essential
Coalitions

Non-essential
Coalitions

saddle Non-saddle

Imperfect info

Symmetric games

Have optimal
equilibrium points
18

Multi-persons

Mixed
strategy

Have no optimal
equilibrium points

fo
.in
rs
de
ea

Mini-Max Search

or

ty

,w

.m

yr

2. The Mini-Max Search Procedure

Consider two players, zero sum, non-random, perfect knowledge games.

ha

kr

ab

Examples: Tic-Tac-Toe, Checkers, Chess, Go, Nim, and Othello.

2.1 Formalizing Game

A general and a Tic-Tac-Toe game in particular.
Consider 2-Person, Zero-Sum, Perfect Information

Both players have access to complete information about the state of

the game.
No information is hidden from either player.
Players alternately move.
Apply Iterative methods

Required because search space may be large for the games to

search for a solution.
Do search before each move to select the next best move.
19

fo
.in
rs
de
ea
yr
or

ty

,w

.m

Mini-Max Search

Adversary Methods
Required because alternate moves are made by an opponent ,

ha

kr

ab

who is trying to win, are not controllable.

Static Evaluation Function

f(n)

Used to evaluate the "goodness" of a configuration of the game.

It estimates board quality leading to win for one player.
Example: Let the board associated with node n then
If f(n) = large +ve value
means the board is good for me and bad for opponent.
If f(n) = large -ve value
means the board is bad for me and good for opponent.
If f(n) near 0
Means the board is a neutral position.
If f(n) = +infinity
means a winning position for me.
If f(n) = -infinity
means a winning position for opponent.
20

fo
.in
rs
yr

ea

de

One player's loss is the other player's gain.

ab

or

ty

,w

.m

Mini-Max Search

Zero-Sum Assumption

ha

kr

Do not know how our opponent plays ?;

So use a single evaluation function to describe the goodness of a

board with respect to both players.
Example : Evaluation Function for the game Tic-Tac-Toe :
f(n) = [number of 3-lengths open for me]

[number of 3-lengths open for opponent]

where a 3-length is a complete row, column, or diagonal.

21

fo
.in
rs
de
ea

Mini-Max Search

or

ty

,w

.m

yr

2.2 MINI-MAX Technique

For Two-Agent , Zero-Sum , Perfect-Information Game.

kr

ab

The Mini-Max procedure can solve the problem if sufficient computational

ha

resources are available.

Elements of Mini-Max technique
Game tree (search tree)
Static evaluation,
e.g., +ve for a win, -ve for a lose and 0 for a draw or neutral.
Backing up the evaluations, level by level, on the basis of
opponent's turn.
22

fo
.in
rs
de
ea
yr
or

ty

,w

.m

Mini-Max Search

Game Trees : description

Root node represents board configuration and decision required

ha

kr

ab

as to what is the best single next move.

If

my

turn

to move, then

the root is labeled a MAX node

indicating it is my turn;
otherwise

it

is

labeled

MIN node

to

indicate

it

is

my

opponent's turn.
Arcs

represent

the possible legal moves for the player that the

arcs emanate from.

At each level, the tree has nodes that are all MAX or all MIN;
Since moves alternate, the nodes at level
from those at level
23

i+1

are of opposite kind

fo
.in
rs
yr

ea

de

Searching Game Tree using the Mini-Max Algorithm

ab

or

ty

,w

.m

Mini-Max Search

Mini-Max Algorithm

ha

kr

Steps used in picking the next move:

Since it's my turn to move, the start node is MAX node with
current board configuration.
Expand nodes down (play) to some depth of look-ahead in the
game.
Apply evaluation function at each of the leaf nodes
"Back up" values for each non-leaf nodes until computed for the
root node.
At MIN nodes, the backed up value is the minimum of the values
associated with its children.
At MAX nodes, the backed up value is the maximum of the
values associated with its children.
Note: The process of "backing up" values gives the optimal strategy,
that is, both players assuming that your opponent is using the
same static evaluation function as you are.
24

fo
.in
rs
de
ea
yr
,w

.m

Mini-Max Search

Example : Mini-Max Algorithm

The MAX player considers all

or

ty

5 S max

ha

kr

ab

three possible moves.

5B

2 C min

-1 A

The opponent MIN player also

considers all possible moves.

D
9

G H

I J

3 -1

6 5

2 7

The evaluation function is applied

to leaf level only.

Apply Evaluation function :

Apply static evaluation function at leaf nodes & begin backing up.
First compute backed-up values at the parents of the leaves.
Node A is a MIN node ie it is the opponent's turn to move.
A's backed-up value is -1 ie min of (9, 3, -1), meaning
if

opponent ever reaches this node,

then it will pick the move associated with the arc from A to F.
Similarly, B's backed-up value is 5 and
C's backed-up value is 2.

Next, backup values to next higher level,

Node S is a MAX node ie it's our turn to move.
look best on backed-up values at each of S's children.
the best child is B since value is 5 ie max of (-1, 5, 2).
So the minimax value for the root node S is 5, and
the move selected is associated with the arc from S to B.
25

fo
.in
rs
de
ea

Game playing with Mini-Max

or

ty

,w

.m

yr

3. Game Playing with Mini-Max - Tic-Tac-Toe

Here, Minimax Game Tree is used that can program computers to play games.

kr

ab

There are two players taking turns to play moves.

ha

Physically, it is just a tree of all possible moves.

3.1 Moves
Start: X's Moves

26

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty
or
ab
kr
ha
C
C

27

Game playing with Mini-Max

Next: O's Moves

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty
or
ab
kr
ha
C
C

28

Game playing with Mini-Max

Again: X's moves

fo
.in
rs
de
ea

Game playing with Mini-Max

+1

for a win,

for a draw

ab

or

ty

,w

.m

yr

3.2 Static Evaluation:

ha

kr

Criteria +1 for a Win, 0 for a Draw

29

fo
.in
rs
de
ea

Game playing with Mini-Max

Level by level, on the basis of opponent's turn

ab

or

ty

,w

.m

yr

3.3 Back-up the Evaluations:

ha

kr

Up : One Level

30

fo
.in
rs
de
ea
yr
.m
w
w
,w
ty
or
ab
kr
ha
C
C

31

Game playing with Mini-Max

Up : Two Levels

fo
.in
rs
de
ea

Game playing with Mini-Max

Choose best move which is maximum

ha

kr

ab

or

ty

,w

.m

yr

3.4 Evaluation obtained :

Best move

32

fo
.in
rs
de
ea

Alpha-beta pruning

or

ty

,w

.m

yr

4. Alpha-Beta Pruning
The problem with Mini-Max algorithm is that the number of game states

ha

kr

ab

it has to examine is exponential in the number of Moves.

The Alpha-Beta Pruning helps to arrive at correct Min-Max algorithm decision

without looking at every node of the game tree.

While using Mini-Max, some situations may arise when search of a particular
branch can be safely terminated. So, while doing search, figure out those nodes
that do not require to be expanded. The method is explained below :

Max-player cuts off search when he knows Min-player can force a provably
bad outcome.

Min player cuts of search when he knows Max-player can force provably
good (for max) outcome

Applying an alpha-cutoff means we stop search of a particular branch

because we see that we already have a better opportunity elsewhere.

Applying a beta-cutoff means we stop search of a particular branch because

we see that the opponent already has a better opportunity elsewhere.

33

Applying both forms is alpha-beta pruning.

fo
.in
rs
de
ea

Alpha-cutoff pruning

or

ty

,w

.m

yr

5.1 Alpha-Cutoff
It may be found that, in the current branch, the opponent can achieve a

kr

ab

state with a lower value for us than one achievable in another branch. So

ha

the current branch is one that we will certainly not move the game to.

Search of this branch can be safely terminated.

34

fo
.in
rs
de
ea

Beta-cutoff

or

ty

,w

.m

yr

5.2 Beta-Cutoff
It is just the reverse of Alpha-Cutoff.

kr

ab

It may also be found, that in the current branch, we would be able to

ha

achieve a state which has a higher value for us than one the opponent can

hold us to in another branch. The current branch can be identified as one

that the opponent will certainly not move the game to.
branch can be safely terminated.

35

Search in this

fo
.in
rs
de
ea
yr
.m
w

AI- Game Playing

or

ty

,w

6. References : Textbooks

kr

ab

1. "Artificial Intelligence", by Elaine Rich and Kevin Knight, (2006), McGraw Hill

ha

companies Inc., Chapter 12, page 305-326.

2. "Artificial Intelligence: A Modern Approach" by Stuart Russell and Peter Norvig,

(2002), Prentice Hall, Chapter 6, page 161-189.

3. "Computational Intelligence: A Logical Approach", by David Poole, Alan Mackworth,

and Randy Goebel, (1998), Oxford University Press, Chapter 4, page 113-163.

4. "AI: A New Synthesis", by Nils J. Nilsson, (1998), Morgan Kaufmann Inc., Chapter
12, Page 195-213.

5. Related documents from open source, mainly internet.

being prepared for inclusion at a later date.

36

An exhaustive list is