摘要
游戏地图自动生成是目前过程内容生成PCG(Procedural Content Generation)研究的热点之一.本文以开源RTS游戏MegaGlest的地图为研究对象,以公平性、可玩性、战略性和趣味性为优化目标,提出多目标粒子群优化的游戏地图生成算法.实验结果表明,自动生成的地图在4个优化目标方面具有明显改进,能给玩家提供更好的游戏体验.
Game map generation is one of the hot spots of procedural content generation (PCG). In this article, the map of MegaGlest which is a open source RTS game is the object of study. Fairness, play- ability, strategy and interesting are regarded as the optimization goals. Multiobjective Particle Swarm Optimization (MOPSO) is presented for game map generation. The results of experiment show that the maps generated have clear improvements in four optimization goals, and can improve the gaming experi- ence of players.
出处
《四川大学学报(自然科学版)》
CAS
CSCD
北大核心
2013年第1期67-72,共6页
Journal of Sichuan University(Natural Science Edition)
基金
四川省科技支撑计划项目(2012GZ0091)
关键词
游戏地图生成
多目标优化
粒子群算法
优化目标
game map generation, multiobjective optimization, particle swarm algorithm, design goal