课程: AI Algorithms for Game Design with Python

免费学习该课程！

今天就开通帐号，24,700 门业界名师课程任您挑！

Depth-limited search

Depth-limited search - Python教程

课程: AI Algorithms for Game Design with Python

开始一个月试用 购买团队方案

Depth-limited search

“

- [Instructor] So let's stop for a moment and ask ourselves, what's stopping us from doing better? Well, complexity. Minimax inherently runs on the depth-first search tree traversal algorithm. So it's really impossible to do the whole thing significantly faster. But wait, the fact that we cannot find a perfect solution to this problem in a short time doesn't mean that we cannot find a reasonable solution in a short time. This has been the key enabler to useful AI we are using today, like traffic apps or self-driving cars. A solution produced in a short amount of time doesn't have to be bad. So with this in mind, we could make some compromises. For example, we could prune further by identifying and skipping states that we've already seen. We could even push this to identifying rotations and reflections of already-seen states. We could also reduce b, the branching factor, by selecting the, say, three most promising moves while leaving unproductive or self-harming moves unchecked. And we…

内容