+ All Categories
Home > Documents > Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are...

Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are...

Date post: 18-Jan-2018
Category:
Upload: beatrix-lester
View: 227 times
Download: 0 times
Share this document with a friend
Description:
Tic-tac-toe XOX X XOO 3 u Two players, max and min u Loop: u Max put X u Min put O u Until one play takes 3 contiguous space or all spaces are filled +1 XOX OX O XOX OOX XXO 0
18
Adversarial Search
Transcript
Page 1: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

Adversarial Search

Page 2: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

Game playing

Multi-agent competitive environment The most common games are deterministic, turn-

taking, two-player, zero-sum game of perfect information

In AI terminology: deterministic, fully observable, two agents act alternately, the utility values at then end are always equal and opposite

2

Page 3: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

Tic-tac-toe

X O XX

X O O

3

Two players, max and min Loop:

Max put X Min put O Until one play takes 3 contiguous space or all spaces are filled

+1

X O XO XO-1

X O XO O XX X O

0

Page 4: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

4

Page 5: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

5

Higher values are good for max,but bad for min

MAX assumes that min plays optimally What if MIN does not play optimally?

Is MAX still optimal?

Page 6: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

6

Page 7: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

7

Page 8: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

8

Page 9: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

9

Page 10: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

10

Page 11: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

11

Page 12: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

12

Page 13: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

Another example

13

max

min

max

4

[4, +∞]

[-∞, 3]

(3)

[-∞, 3]

4

Page 14: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

14

Assuming a left-to-right evaluation of the tree, give the most general relation between x, y, z, w, k, l and m that must hold such that the tree in the dotted part would be pruned by alpha-beta pruning

Page 15: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

15

Page 16: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

Move ordering

Best ordering leads to O(bm/2) time complexity

16

Page 17: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

17

The perfect order is not known until all child nodes are exploredRandomized order works reasonably well

Page 18: Adversarial Search. Game playing u Multi-agent competitive environment u The most common games are deterministic, turn- taking, two-player, zero-sum game.

Summary Deterministic, turn-taking, two-player, zero-sum

game of perfect information Minimax algorithm, optimal for optimal player Not all nodes need to be explored, alpha-beta

pruning

18

HW1 due next Monday,Exam 1 on next Monday


Recommended