Novelty search for evolving interesting character mechanics for a two-player video game
Eirik Høgdahl Skjærseth, Harald Vinje, Ole J. Mengshoel · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2021
Computational generation of video game content, often referred to as procedural content generation (PCG), holds much promise for generating character mechanics. Character mechanics refers to how characters are allowed to move and behave in a computer game, rather than aesthetics such as graphics and audio. In this paper we study how to generate character mechanics automatically, by means of novelty search. Our results show that some of the auto-generated characters are, by human subjects, perceived as more interesting than built-in game characters.