Design of Emergent and Adaptive Virtual Players in a War RTS Game - Foundations on Natural and Artificial Computation - page 371

Design of Emergent and Adaptive Virtual
Players in a War RTS Game

Jos´eA.Garc ıa Gutierrez, Carlos Cotta, and Antonio J. Fernandez Leiva

Dept. Lenguajes y Ciencias de la Computacion, ETSI Informatica,
Campus de Teatinos, Universidad de Malaga,
29071 Malaga - Spain
{ ccottap,afdez } @lcc.uma.es

Abstract. Basically, in (one-player) war Real Time Strategy (wRTS)
games a human player controls, in real time, an army consisting of a num-
ber of soldiers and her aim is to destroy the opponent's assets where the
opponent is a virtual (i.e., non-human player controlled) player that usu-
ally consists of a pre-programmed decision-making script. These scripts
have usually associated some well-known problems (e.g., predictability,
non-rationality, repetitive behaviors, and sensation of artificial stupidity
among others). This paper describes a method for the automatic gener-
ation of virtual players that adapt to the player skills; this is done by
building initially a model of the player behavior in real time during the
game, and further evolving the virtual player via this model in-between
two games. The paper also shows preliminary results obtained on a one-
player wRTS game constructed specifically for experimentation.

1 Introduction and Related Work

In an era in which computing power has boosted the graphical quality of
videogames, players have turned their attention to other aspects of the game. In
particular, they mostly request opponents exhibiting intelligent behavior. How-
ever, intelligence does not generally equate to playing proficiency but rather to
interesting behaviors [1]. This issue is specifically relevant in real-time strategy
(RTS) games that often employ two kinds of artificial intelligence (AI) [2]: one
represented by a virtual player (VP, or non-player character - NPC) making
decisions on a set of units (i.e., warriors, machines, etc.), and another one cor-
responding to the small units (usually with little or no intelligence). The design
of these AIs is a complex task, and the reality in the industry is that in most of
the RTS games, the NPC is basically controlled by a fixed script that has been
previously programmed based on the experience of the designer/programmer.
This script often comprises hundreds of rules, in the form if the game is in

state S then Unit U should execute action A , to control the behavior of

the components (i.e., units) under its control. This arises known problems: for
instance, for players, the opponent behavior guided by fixed scripts can become
predictable for the experienced player. Also, for AI programmers, the design of
virtual players (VPs) can be frustrating because the games are becoming more