Transcript Lecture 6
COIT23003 Games Development 8. Elaboration: Behaviour References • Meigs, T., 2003. Ultimate Game Design: Building Game Worlds, McGraw Hill. • Fairclough, C. et al, 2001. Research Directions for AI in Computer Games http://www.comp.dit.ie/bmacnamee/papers/ ResearchDirectionsForGameAI.pdf • McGugan, W., 2007. Beginning Game Development with Python and Pygame: From Novice to Professional. Apress. Behaviour • Entity behaviour is critical for providing an interesting/realistic/challenging gaming experience • Two major approaches: – Scripting – Artificial Intelligence (AI) Scripted Behaviour • Entity behaviours are predefined and hard coded, (ie scripted) eg If (monster.canSeePlayer) Monster.attack(); Else if (monster.canSeeWeapon() … • Not to be confused with scripting languages Scripting Languages • Scripting languages are languages that are – (Perceived to be) simpler to use (eg JavaScript, Python, Lua, …) than “real” languages (C++, Java, etc) and/or – Used to control applications (eg DOS Batch Language, Visual Basic) or libraries (eg Python) • Games development typically employs scripting languages because of their “simplicity”. Typically, the whole game is developed in the scripting language, although languages like Python allow you to access libraries developed in other languages. Scripting Languages • Eg Pygame is a wrapper for the SDL library (written in C/C++) Class Discussion • What do you think might be the pluses/minuses in developing games using Python/Pygame vs C++/SDL? Scripted Behaviour • Scripted behaviours can occur either through – Entity deliberation (what should I do next) or – Environmental triggers (what should I do in response to in response to event x) • The two approaches are not exclusive – triggers can result in deliberation (I’ve been shot at, what do I do?). The issue is that with scripting, the choices have been determined (and hard coded) by the developer Scripted Behaviour: Summary • Introducing variability into entity behaviours can be challenging • Scripted behaviour is simple and straightforward to implement Intelligent Behaviour • Intelligent behaviour (among other things) requires the ability to plan and to learn. The routine implementation of these abilities is a long way off, so they are rarely seen in computer games. Rather, what we do see is the incorporation of simple and robust techniques from AI that make it easier for developers to produce entities that appear to behave intelligently. AI Techniques Examples of the use of commonly used AI techniques include • Specialised search techniques (eg A*) for route planning • Finite state machines for structuring entity behaviour • Flocking algorithms for coordinated group movements Implementation • For an example of simple state machine implementation, refer to McGugan, Chapter 7. • Discussion question: – Is there any fundamental difference between using state machines or scripting to represent behaviour? i.e. are entities that employ state driven behaviour any more “intelligent” than entities that employ scripting? Applications • Game AI is different to conventional applied AI – the objective is to produce realistic and interesting behaviour as opposed to solving difficult problems. Hence AI might be used to produce – Challenging, but not invincible entities – “Dumb” entities – Entities that predict what the player might do next (Fairclough et al, 2007) Intelligent Behaviour: Summary • Focus remains on simple, robust techniques (Fairclough et al, 2007) • Provision of “real” intelligence remains a research issue A Design Example • Meigs, pp. 154-159 • Discussion Question: • Variability is introduced through random number generation. Do you think that this, by itself, would provide for interesting gameplay? How might you improve the gameplay experience?