Game Description Language (GDL), an innovation in the field of artificial intelligence[citation needed], represents a significant step forward in the quest to create versatile game-playing AI agents. Designed by Michael Genesereth, GDL is a specialized logic programming language that finds its home within the ambitious realm of the General Game Playing Project at Stanford University. This project aims to develop AI agents capable of playing a wide variety of games without the need for game-specific programming. At its core, GDL serves as a powerful tool for expressing the intricacies of game rules and dynamics in a form that is easily comprehensible to AI systems. It achieves this through a combination of logic-based constructs and declarative principles.
Here are key features and concepts associated with Game Description Language:
**Logic-Based Foundations:** GDL leverages formal logic to describe the rules of games. It allows for the precise representation of complex game mechanics, providing unambiguous instructions for gameplay. By using logical rules, GDL enables AI agents to understand and navigate the intricacies of a game's structure and dynamics.
**Game-Agnostic Flexibility:** A notable strength of GDL is its game-agnostic nature. It is not confined to any specific game or genre, making it a versatile language for defining the rules of an extensive range of games. This versatility is particularly valuable in the development of AI agents, as it empowers them to engage in various games without the need for individualized programming for each one.
**Declarative Clarity:** GDL adheres to a declarative paradigm. In other words, it focuses on defining the conditions that are either true or legal within the context of a game, rather than prescribing how to play the game. This clear distinction between rules and strategies is vital for AI agents. It enables them to reason logically about the game state and make informed, intelligent moves.
**Extensible Adaptability:** Game Description Language is an extensible framework. It can be expanded and customized to accommodate a wide array of game features. Whether a game involves chance elements, simultaneous actions, or other unique mechanics, GDL can be adapted to accurately represent these complexities.
In practice, GDL serves as a cornerstone for the General Game Playing competitions and research endeavors. In these contexts, GDL is used to specify the rules of games that AI agents are expected to play. AI developers and researchers harness GDL to create algorithms that can comprehend and engage with games based on their rule descriptions. The use of GDL paves the way for the development of highly adaptable AI agents, capable of competing and excelling in diverse gaming scenarios.
This innovation is a testament to the convergence of logic-based formalism and the world of games, opening new horizons for AI's potential in understanding and mastering a multitude of games. Game Description Language equips AI with a universal key to unlock the mysteries of diverse game environments and strategies.
Quoted in an article in New Scientist, Genesereth pointed out that although Deep Blue can play chess at a grandmaster level, it is incapable of playing checkers at all because it is a specialized game player.[1] Both chess and checkers can be described in GDL. This enables general game players to be built that can play both of these games and any other game that can be described using GDL.
GDL is a variant of Datalog, and the syntax is largely the same. It is usually given in prefix notation. Variables begin with "?
".[2]
The following is the list of keywords in GDL, along with brief descriptions of their functions:
distinct
does
does(?r,?m)
means that player (or role) ?r
makes move ?m
in the current game state.goal
goal(?r,?n)
is used to define goal value ?n
(usually a natural number between 0 and 100) for role ?r
in the current state.init
legal
legal(?r,?m)
means that ?m
is a legal move for role ?r
in the current state.next
role
terminal
true
A game description in GDL provides complete rules for each of the following elements of a game.
Facts that define the roles in a game. The following example is from a GDL description of the two-player game Tic-tac-toe:
(role xplayer) (role oplayer)
Rules that entail all facts about the initial game state. An example is:
(init (cell 1 1 blank)) ... (init (cell 3 3 blank)) (init (control xplayer))
Rules that describe each move by the conditions on the current position under which it can be taken by a player. An example is:
(<= (legal ?player (mark ?m ?n)) (true (cell ?m ?n blank)) (true (control ?player)))
Rules that describe all facts about the next state relative to the current state and the moves taken by the players. An example is:
(<= (next (cell ?m ?n x)) (does xplayer (mark ?m ?n))) (<= (next (cell ?m ?n o)) (does oplayer (mark ?m ?n)))
Rules that describe the conditions under which the current state is a terminal one. An example is:
(<= terminal (line x)) (<= terminal (line o)) (<= terminal not boardopen)
The goal values for each player in a terminal state. An example is:
(<= (goal xplayer 100) (line x)) (<= (goal oplayer 0) (line x))
With GDL, one can describe finite games with an arbitrary numbers of players. However, GDL cannot describe games which contain an element of chance (for example, rolling dice) or games where players have incomplete information about the current state of the game (for example, in many card games the opponents' cards are not visible). GDL-II, the Game Description Language for Incomplete Information Games, extends GDL by two keywords that allow for the description of elements of chance and incomplete information:[3]
sees
sees(?r,?p)
means that role ?r
perceives ?p
in the next game state.random
The following is an example from a GDL-II description of the card game Texas hold 'em:
(<= (sees ?player ?card) (does random (deal_face_down ?player ?card))) (<= (sees ?r ?card) (role ?r) (does random (deal_river ?card)))
Michael Thielscher also created a further extension, GDL-III, a general game description language with imperfect information and introspection, that supports the specification of epistemic games — ones characterised by rules that depend on the knowledge of players.[4]
In classical game theory, games can be formalised in extensive and normal forms. For cooperative game theory, games are represented using characteristic functions. Some subclasses of games allow special representations in smaller sizes also known as succinct games. Some of the newer developments of formalisms and languages for representation of some subclasses of games or representations adjusted to the needs of interdisciplinary research are summarized as the following table.[5] Some of these alternative representations also encode time related aspects:
Name | Year | Means | Type of games | Time |
---|---|---|---|---|
1973 | functions | subset of n-person games, simultaneous moves | No | |
1994 | matrices | 2-person games of imperfect information | No | |
Timed games[6][7] | 1994 | functions | 2-person games | Yes |
Gala[8] | 1997 | logic | n-person games of imperfect information | No |
Graphical games[9][10] | 2001 | graphs, functions | n-person games, simultaneous moves | No |
Local effect games[11] | 2003 | functions | subset of n-person games, simultaneous moves | No |
Game Petri-nets[12] | 2006 | Petri net | deterministic n-person games, simultaneous moves | No |
Continuous games[13] | 2007 | functions | subset of 2-person games of imperfect information | Yes |
PNSI[14][15] | 2008 | Petri net | n-person games of imperfect information | Yes |
Action graph games[16] | 2012 | graphs, functions | n-person games, simultaneous moves | No |
A 2016 paper "describes a multilevel algorithm compiling a general game description in GDL into an optimized reasoner in a low level language".[17]
A 2017 paper uses GDL to model the process of mediating a resolution to a dispute between two parties, and presented an algorithm that uses available information efficiently to do so.[18]
Original source: https://en.wikipedia.org/wiki/Game Description Language.
Read more |