Automatic generation of game elements via evolution
Automatic generation of game elements via evolution
复制标题
通过进化自动生成游戏元素
DOI：
10.1109/itw.2010.5593341
复制
发表时间：
2010
期刊：
影响因子：
--
通讯作者：
D. Ashlock
中科院分区：
文献类型：
--
作者：
D. Ashlock
本研究提出了一个系统，用于自动产生拼图游戏设计中使用。该系统采用了一种进化算法，将谜题优化到指定的难度水平。适应度函数使用动态规划来计算解决难题所需的最小移动次数。探索了两种类型的谜题，一种是基于棋子的迷宫，另一种是使用颜色相关的色轮来创建隐式迷宫。进化算法能够在指定的难度水平上产生各种各样的谜题。因此，该算法可以用于提供用于游戏设计或用于可变游戏内容的库。该技术是灵活的，可以推广到显着的复杂性，通过简单地升级的动态规划算法中使用的适应度函数的难题。结果发现，需要最大数量的移动来解决的难题，潜在的，不太难，因为这样的难题，目前的球员很少的选择。这个问题是通过修改算法来解决的，以搜索具有较小的最小移动数的谜题，在谜题中留下更多的选择空间和随之而来的混乱。
This study presents a system for automatically producing puzzles for use in game design. The system incorporates an evolutionary algorithm that optimizes the puzzle to a specified level of difficulty. The fitness function uses dynamic programming to compute the minimum number of moves required to solve a puzzle. Two types of puzzle are explored, one is a maze based on chess pieces and the other uses colors as related by the color wheel to create an implicit maze. The evolutionary algorithm is able to produce a wide variety of puzzles at specified levels of difficulty. The algorithm can thus be used to provides a library for game design or for variable game content. The technique is flexible and can be generalized to puzzles of remarkable complexity by simply upgrading the dynamic programming algorithm used in the fitness function. It is found that puzzles requiring a maximum number of moves to solve are, potentially, less difficult because such puzzles present the player with few choices. This problem is addressed by modifying the algorithm to search for puzzles with a smaller minimum number of moves required, leaving more room in the puzzle for choice and its attendant confusion.