Insights Blog
-- Browse All Articles --
Physics Articles
Physics Tutorials
Physics Guides
Physics FAQ
Math Articles
Math Tutorials
Math Guides
Math FAQ
Education Articles
Education Guides
Bio/Chem Articles
Technology Guides
Computer Science Tutorials
Forums
Chemistry
Biology and Medical
Earth Sciences
Computer Science
Computing and Technology
DIY Projects
Trending
Featured Threads
Log in
Register
What's new
Search
Search
Search titles only
By:
Chemistry
Biology and Medical
Earth Sciences
Computer Science
Computing and Technology
DIY Projects
Menu
Log in
Register
Navigation
More options
Contact us
Close Menu
JavaScript is disabled. For a better experience, please enable JavaScript in your browser before proceeding.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an
alternative browser
.
Forums
Other Sciences
Programming and Computer Science
Why doesn't Pac-Man eat the dot? Berkeley's AI course 188
Reply to thread
Message
[QUOTE="Korisnik, post: 6195571, member: 427675"] Thank you for your reply. Do you see the example the professor shows next, where there's a very simple one-row "game space" (this blue environment Pacman moves in) of only a few possible spaces, with a dot in each corner, left and right. [ATTACH type="full" alt="245379"]245379[/ATTACH] So, this is a different scenario where at the beginning Pacman plans 2 levels ahead. It's obvious here that it can happen that it simply follows the first right arm of the tree, but then it replans, which means it gets this same tree (with subtrees symmetrically flipped): [ATTACH type="full" alt="245381"]245381[/ATTACH] So now because the minimax algorithm gives it the same values for both subtrees, it can again reasonably take the left subtree. And so on, it paces back and forth, never eating the dot.In the situation I described above, this doesn't happen. I have sketched the tree on a paper for each step of the planning. [ATTACH type="full" alt="245382"]245382[/ATTACH] So, it first plans and takes the A route (it could've taken B just as well), which puts it above the dot. Then it plans again, and in step 2 it should take B because it can get the most out of it. Yes, it can see the other dot, and I suspect there might be something going on with it taking it into account, but as shown in the other example, it doesn't seem like it considers it at all; only the [B]score[/B]. Yes, this is to show that replanning can sometimes lead to starvation (or thrashing, or whatever it's called), but I'm just not sure how it works on that particular example. [/QUOTE]
Insert quotes…
Post reply
Forums
Other Sciences
Programming and Computer Science
Why doesn't Pac-Man eat the dot? Berkeley's AI course 188
Back
Top