Transcript Global Domination -RISK-

Global Domination
-RISK-
Table of Contents



Introduction
 Type Of Game
 Formats
 Game Play
Artificial Intelligence
 Type of A.I.
 Variations of the A.I.
Improvements to the A.I.
Introduction
•
•
•
Type of game
History of the game
Platforms
Type of game
Turn based strategy game
 Multiple computer players available



Choices range 2 – 6 players
Multiple variations of the rules
Set number of troops for “turning in a book”
 Increasing number of troops for “turning in a
book”

History
Created by Hasbro
 Originally a board game
 First computer version was available with
Commodore 64, and IBM compatible
machines



B.P. (Before Pentium)
Many different versions exist for
marketability (Lord of the Rings, Star
Wars, etc…)
Platforms

PC
400mhz+ for good game play
 Windows ’95 or greater

 Slight
problems with Windows XP
Mac
 Play Station 2
 Evaluation Version: PC

Game Play




Initial number of troops to set on a random deal of
countries
Each turn you get reinforcements based on the number
of countries you control, and the continents you control
Cards are given at the end of the turn if a successful
conquest took place
 Cards will have a cannon, Knight, or soldier on it.
 A book can be three of a kind, or one of each
Goal is to:
 Overpower opponents
 Defeat them from the game
 Conquer the world
Artificial Intelligence

Risk utilizes prediction algorithms and
battle strategies
Actual algorithms were not disclosed
 Popular heuristic algorithm were more than
likely used within a powerful rule based
system for the battle strategy
 Rule Based systems often use human experts
as guides

Different Algorithms

Exhaustive Search
Good for simple games (Tic-Tac-Toe)
 More than likely not used for Risked based on
the number of combinations and decisions to
take into account

Improved Searches
Main goal of a prediction algorithm is to
make an informed search to locate as few
open states as possible based on the
search diagram
 Such Heuristic searches can be:

Best-First (hill climbing)
 A*
 Minimax

Best First Search
Expand the current node
 Evaluate its children
 Moves onto best child for further
expansion

A*
If algorithm A is used with an evaluation
function in which h(n) <= to the cost of the
minimal path from n to the goal
 Given a search function h(n)
 Known starting point and goal
 A* will produce the shortest path


For game design, these parameters may not
be known
Minimax
Assumes that the opponent or opponents
use the same knowledge of the play
space, and uses that knowledge to win the
game
 Opponents are referred to as Min and Max

Max represents player trying to win
 Min is the opponent trying to minimize Max’s
score

Minimax cont…
Most logical course of action for any
strategy game
 When incorporated into a battle strategy
system

Can predict to a point a player’s next move.
 Improvements to the searching algorithm and
other variances can create a more difficult A.I.

Battle Strategies
Rule based
 Situation Based
 Game of this level would have an
extensive system

Improvements

Improving both the search algorithms and
case bases
Results in different levels of A.I.
 Results in a superior game with each new
release

This version

By allowing the user to change the way
the over all game is played

Adjusts the variables for both the heuristic
searches and battle strategies