Adaptive AI Engine for RTS Games

Discussing the theory and practice

Archive for the ‘SRSs’ Category

Contains Software Requirements Specification Documents developed throw year

Enhanced AI Engine

Posted by Ogail on June 26, 2009

Enhanced AI Engine

  • Problems Considered:

    • Opponents weaknesses easily detected
      • In most commercial complex major games, human players only need a couple of games to spot opponents’ weaknesses and to exploit them in upcoming games. This results in not effective boredom game
    • Sharing experience is complex
      • A Human/Computer player would like to share his game plan/experience with other players among the world without really playing the game with them
    • The need for tangible experience
      • A Human Player would be interested to read a human-comprehendible game plan of another Computer/Human Player, He Can’t get what he wants from a computer because it’s something embedded in the AI Engine of the game, also he can’t get what he wants from a human player if he isn’t in contact with him
    • Discovering and solving flaws in game plan/experience
      • A Human/Computer player tends to discover his mistakes that lead to his loss, this process may take a long time and effort
  • Alterative Solutions:
    • Opponents weaknesses easily detected:
      • Human players/Testers around the world supply Game Manufacturers with the weak points in the AI algorithms in their games, so they would avoid them in the next version of the game
    • Sharing experience is complex
      • No Other way than playing the game either online or offline
    • The need for tangible experience
      • A Human would contact the game developers and ask them for the computer plans for a certain part of the game, this is extremely inconvenient
      • A Human would wait until he’s able to contact the other human, and have with him a long chat about his experience in the game
    • Discovering and solving flaws in game plan/experience
      • The Human Player would contact the game developer and ask for help, or the game developer would update the computer AI game plan after discovering a flaw in it
  • Objective:
    • Implementing machine learning techniques at AI Engines
      • Solves the first problem
      • Description:
        • Makes the AI Engine able to analyze its failure and solve the flaw
      • Other:
        • The following research were done in this field before:
  • Implementing Human Understandable Experience
    • Solves second problem
    • Description:
      • So that a Human would easily understand the plan of a certain computer opponent or even the electronic version of the game plan of a human player WITHOUT playing the game itself
  • Implementing mobile experience/plan concept
    • Solves third and third problems
    • Description:
      • A Computer/Human player who would like to export his way/Experience of playing the game to other Computer/Human players around the world will share his mobile copy of experience/plan, Other player’s game instances will receive the mobile copy of his experience/plan in the game, perhaps add it to their own copy of way/experience of playing the game and try to beat it
  • Detecting player personality:
    • Solves first problem
    • Description:
      • The AI Engine will be able to analyze player behavior and produces a plan suitable for beating him/her
  • Scope:
    • Our current scope is strategy games
  • Why this project is useful?
    • Currently there are automated machines (robots) that are used in wars instead of human. The army commander can use the AI Engine to:
      • Test his plans
      • Share other commanders experience in the plan
      • Discover flaws in the plan. And solves these flaws
  • Motivations:
    • We are interested in this project for the following reasons:
      • AI is our area of interest in computer science
      • In the future, we are planning to work in the industry of game development specially in AI Engines
      • We love playing strategy games, developing this project will be so interesting and fun!

Our wish to add something new to this world

Posted in SRSs | Leave a Comment »

Our Initial Imagination of the Project – 14 June 2006

Posted by merothehero on June 25, 2009

AI Engine with Machine Learning and Mobile Game Plans

Problems:

1- In most commercial complex major games, Human players only need a couple of games to spot opponents’ weaknesses and to exploit them in upcoming games. This results in game boredom.

Traditional Solution:

-Human players/Testers around the world supply Game Manufacturers with the weak points in the AI algorithms in their games, so they would avoid them in the next version of the game.

2- A Human/Computer player would like to share his game plan/experience with other players among the world without really playing the game with them.

Traditional Solution:

-No Other way than playing the game.

3- -A Human Player would be interested to read a human-comprehendible game plan of another Computer/Human Player, He Can’t get what he wants from a computer because it’s something embedded in the AI Engine of the game, also he can’t get what he wants from a human player if he isn’t in contact with him.

Traditional Solutions:

-A Human would contact the game developers and ask them for the computer plans for a certain part of the game, this is extremely inconvenient.

-A Human would wait until he’s able to contact the other human, and have with him a long chat about his experience in the game.

Objective

1- Implementing Machine Learning in AI engines.

In most of Commercial Complex Major Games, Machine Learning isn’t implemented. Implementing Machine learning in an AI game engine would be great.

Please check this research:

Recent Advances in Machine Learning and Game Playing (Research between 2006 and 2009)

http://www.ke.tu-darmstadt.de/~juffi/publications/ogai-07.pdf

Another Link: http://209.85.129.132/search?q=cache:dxP3xXir7mkJ:www.cs.ualberta.ca/~mburo/ps/RTS-ijcai03.ps.gz+machine+learning+strategy+games&cd=7&hl=en&ct=clnk&gl=eg

2- Implementing a mobile experience/plan

A Computer/Human player who would like to export his way/Experience of playing the game to other Computer/Human players around the world will share his mobile copy of experience/plan, Other player’s game instances will receive the mobile copy of his experience/plan in the game, perhaps add it to their own copy of way/experience of playing the game and try to beat it.

3- Implementing Human Understandable Experience

So that a Human would easily understand the plan of a certain computer opponent or even the electronic version of the game plan of a human player WITHOUT playing the game itself.

Motivations

The World is heading through a new era of gaming and the game industry will need this sooner or later.

Scope

Our current scope is Strategy Games.

Posted in SRSs | Leave a Comment »