A First Study on Genetic Algorithms Based-Evolvable Motivation Model for Virtual Agents

Abstract

In video games, it's very important to model a virtual agent with motivation variables. In previous researches, a virtual agent's motivation variables are always static. An evolvable motivation system for a virtual agent is set up, priority sequence of an agent's behaviors is decided by the priority sequence of its motivations, the priority sequences of an agent 's motivations are determined by genetic algorithms, the demo system shows that virtual agents can adapt to the environment better.