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！

Is alpha-beta pruning still relevant?

Is alpha-beta pruning still relevant? - 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

Is alpha-beta pruning still relevant?

“

- [Instructor] So now that we have depth limited search to pretty much take as long as we can afford to respond, is alpha-beta pruning still relevant? After all, recall how unimpressive the improvement of alpha-beta pruning was. At best, it reduces the exponent to its half, meaning that it reduces the running time to its square root. To put this in perspective, it would reduce 1 billion years to about 32,000 years. That's not so impressive. However, recall that splitting the exponent in half also means computing twice the levels in the same amount of time. So in the best case, if we use regular min and max with depth limited search on a tree of some depth N, and it takes, say, one minute to compute, then by simply adding alpha-beta pruning, we could compute a tree of twice its depth in that same minute. So the upshot is that we get to compute deeper trees. And why is this good? Well, because evaluation functions are usually far from accurate, and their accuracy improves with depth. It…

内容