At the moment, we have tags dynamic-programming and dynamic-optimization. Should they be merged?
$\begingroup$
$\endgroup$
5
-
$\begingroup$ Dynamic programming can have also second meaning when we talk about computer programming and while most of such questions belong to StackOverflow as far as I understand if the problem connects to economics it would be still on topic here $\endgroup$– 1muflon1 ModJul 25, 2020 at 12:36
-
$\begingroup$ @1muflon1 What 2nd meaning is that? In computer programming, DP has a discrete flavor. It's still the same field. Related: math.stackexchange.com/a/2819083 $\endgroup$– Rodrigo de AzevedoJul 25, 2020 at 12:56
-
$\begingroup$ @1muflon1 For example, take a look at slides 16-20 of this lecture from Stanford EE363. $\endgroup$– Rodrigo de AzevedoJul 25, 2020 at 13:07
-
$\begingroup$ the meaning is also similar, dynamic programming would be writing some program solving optimal substructure or overlapping subproblem. Still it has distinct flavor as you put it- they are not just mere synonyms. I think that justifies having separate tags, is there any advantage of having less granular tags anyway? $\endgroup$– 1muflon1 ModJul 25, 2020 at 13:09
-
$\begingroup$ @1muflon1 Usually, I am all for more granularity. However, in this case, I would advocate for less. The distinction seems to be modern. I would dare to say that Richard Bellman himself would have found it strange. After all, what is the name of his book? I find the distinction anti-historical. The principle is the same in the discrete and continuous settings. However, in the discrete case, the principle itself gets sort of conflated with the details pertaining to the implementation in software. Since this is not a software forum, I find the distinction unnecessary. $\endgroup$– Rodrigo de AzevedoJul 25, 2020 at 14:18
Add a comment
|