Content deleted Content added
No edit summary |
LouScheffer (talk | contribs) →Two kinds of galactic-ness?: new section |
||
Line 29:
The example given of simulated annealing with a logarithmic cooling schedule does not seem to fit the definition of a galactic algorithm as being asymptotically better than practical solutions, unless I am misunderstanding?
[[Special:Contributions/2A00:23C6:4C28:9E01:E9CC:320E:F1BE:BFA4|2A00:23C6:4C28:9E01:E9CC:320E:F1BE:BFA4]] ([[User talk:2A00:23C6:4C28:9E01:E9CC:320E:F1BE:BFA4|talk]]) 16:50, 4 September 2022 (UTC)
== Two kinds of galactic-ness? ==
This article mixes up two closely related kinds of galactic. In both cases, the algorithm is better than any used in practice.
*In one case (like multiplication), it only excels on impractically large examples.
*In other cases (Shannon, Simulated annealing) it gives superior results, but only if impractical amounts of computer power are used.
The second kind was not included in the original definition, but my opinion is that it is in the same spirit - it's a theoretically great algorithm that is not practical. And it has the same benefits as the first kind (inspiring research, showing limits, etc.)
Perhaps a better title might be "Superior but impractical algorithms", but I think "Galactic" is a catchier shorthand for this. [[User:LouScheffer|LouScheffer]] ([[User talk:LouScheffer|talk]]) 20:26, 4 September 2022 (UTC)
|