Artificial intelligence in video games

Artificial intelligence in video games comprises computational techniques that control nonplayer characters, adapt gameplay, generate content, or produce other intelligent behavior within a game.

Finite-state machine: A computational model that switches among a finite set of states according to rules and inputs. Game characters often use states such as patrolling, pursuing, and attacking to organize behavior.
Non-player character: A game character whose actions are controlled by the game rather than directly by a player. NPC behavior is the most visible destination for game AI techniques.
Artificial intelligence: The study and construction of computational systems that perform tasks associated with intelligent behavior. Video-game AI is a specialized application of this broader field.
Rule-based system: A system that applies explicit rules to facts or inputs to produce decisions or outputs. Hand-authored rules remain common beside learned policies in shipped games.
Believable agent: An artificial character whose behavior appears coherent and intentional to human observers. Believability is often a more practical goal for game AI than optimal play.
Behavior tree: A hierarchical structure of tasks and conditions used to select an agent's actions. Behavior trees let game developers combine reusable actions into readable character logic.
Dynamic game difficulty balancing: The adjustment of a game's challenge in response to a player's performance or state. AI can tune encounters or assistance to keep play within a desired difficulty range.
Pathfinding: The process of finding a route between locations through a graph or other representation of space. Many game agents need routes through levels before they can act convincingly.
Machine learning: A field of methods that fit models from data to perform tasks or make predictions. Many game behaviors are scripted rather than learned from data.
Procedural content generation in games: The algorithmic production of playable game content such as levels, maps, or items. Automated generation can expand content while changing how designers shape experiences.

Linked from 6 pages