Skip to main content
14 events
when toggle format what by license comment
Sep 25, 2024 at 5:37 history edited ydd CC BY-SA 4.0
Fixed an issue in the definition of `entropyAsymptotic2`
Sep 24, 2024 at 15:59 comment added Daniel Lichtblau As I note in my own response, integration only is good for the log and constant terms of the asymptotic approximation. So complexity notwithstanding, your method does better.
Sep 23, 2024 at 17:38 comment added ydd @DanielLichtblau I look forward to seeing the answer. Note the IntegrateChangeVariables with t==Sqrt[-1+k] is for the exact summand. If you use asymSummand you can just integrate directly and the asymptotic entropy is very straightforward from there compared to my messy solution
Sep 23, 2024 at 17:36 comment added Daniel Lichtblau I'll post it in a bit. Different methods can be useful to have in answers.
Sep 23, 2024 at 17:05 comment added ydd @DanielLichtblau Ah yes, this does work! Especially if you IntegrateChangeVariables with t==Sqrt[-1+k] . I can add it to the answer if you want, but if you want to post it as your own answer that would be cool (since it was your idea).
Sep 23, 2024 at 16:40 comment added Daniel Lichtblau Maybe a simpler path to the same result: replace the sum by an integral. Then justify that the difference is o(1) as r->infinity. (Roughly: denominator of each term ->0 while numerators change more slowly in r.)
Sep 23, 2024 at 5:28 comment added ydd @A.Kato thank you
Sep 23, 2024 at 4:50 vote accept Arbuja
Sep 23, 2024 at 4:03 history edited ydd CC BY-SA 4.0
added 177 characters in body
Sep 23, 2024 at 3:52 history edited ydd CC BY-SA 4.0
added 192 characters in body
Sep 23, 2024 at 3:45 history edited ydd CC BY-SA 4.0
Removed Abs from plot comparison of values that don't go to zero with r
Sep 23, 2024 at 3:42 comment added A. Kato (+1) Great answer like this again!
Sep 23, 2024 at 3:35 history edited ydd CC BY-SA 4.0
added 859 characters in body
Sep 23, 2024 at 3:09 history answered ydd CC BY-SA 4.0