Skip to main content

Timeline for answer to NSolve for high degree univariate polynomials by Michael E2

Current License: CC BY-SA 3.0

Post Revisions

7 events
when toggle format what by license comment
Apr 13, 2017 at 12:55 history edited CommunityBot
replaced http://mathematica.stackexchange.com/ with https://mathematica.stackexchange.com/
Jan 27, 2016 at 20:32 history edited Michael E2 CC BY-SA 3.0
deleted 227 characters in body
Jun 23, 2014 at 23:44 comment added Michael E2 @John I feel foolish: See the update. You can use Solve on the infinite-precision equation to verify the solutions -- or even find them! You could make a plot near an approximate zero. For example, With[{x0 = N[x /. solexact[[57]], 200]}, With[{eps = 2^Floor@Log2[Abs[$MachineEpsilon x0]]}, ListLinePlot[{Table[{k, Re@eP /. x -> x0 + k eps}, {k, -2, 2, 1/5}], Table[{k, Im@eP /. x -> x0 + k eps}, {k, -2, 2, 1/5}]}, ImageSize -> {Automatic, 150}]]]. Then you can see where it crosses.
Jun 23, 2014 at 23:02 history edited Michael E2 CC BY-SA 3.0
Added solution, clarified explanation
Jun 23, 2014 at 21:34 comment added John Thanks a lot for your so helpful and detailed answer. However, the equation is not zero and that's what is worrying me! Well, let me phrase my worry: how do I make sure that the numerical solutions are indeed near actual solutions? I guess I must use some sort of certification here! In that case, probably it is not a thing to discuss in this forum?
Jun 22, 2014 at 3:18 history edited Michael E2 CC BY-SA 3.0
Improved formatting
Jun 21, 2014 at 23:52 history answered Michael E2 CC BY-SA 3.0