tag:blogger.com,1999:blog-8288194986820249216.post7061840241838334255..comments2024-03-22T05:09:17.789-07:00Comments on Abstract Heresies: More about that puzzleJoe Marshallhttp://www.blogger.com/profile/03233353484280456977noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-8288194986820249216.post-43184618997515672172011-03-11T00:45:40.324-08:002011-03-11T00:45:40.324-08:00The actions could be, for example:
- option 1:
-- ...The actions could be, for example:<br />- option 1:<br />-- wait N seconds and poll,<br />-- give up,<br />- option 2:<br />-- wait 1 second,<br />-- poll,<br />-- give up.<br /><br />Note that, if you don't know the distribution, this can be modeled as a Reinforcement Learning problem (for which an ideal solution would optimally trade off between exploring to acquire more knowledge and exploiting to make more money).<br /><br />Let me add that this interesting puzzle is very similar to the question:<br />"When should I open my oven to check whether the cake is cooked or not?"<br />(with the added difficulty that opening the oven may alter the cooking process)Unknownhttps://www.blogger.com/profile/10827030911542357153noreply@blogger.comtag:blogger.com,1999:blog-8288194986820249216.post-25862435941285396332011-03-08T11:47:42.198-08:002011-03-08T11:47:42.198-08:00If You know the distribution, getting the optimal ...If You know the distribution, getting the optimal payout is a simple dynamic programming problem.<br /><br />the state would be (curent time, time of last poll).marcinhttps://www.blogger.com/profile/17086753413746106828noreply@blogger.com