Skip to content

Integrate a power of a sine or hyperbolic sine over a binomial in a power of it over its roots - #1911

Merged
Rafael-SOWNet merged 4 commits into
masterfrom
a-power-over-a-binomial-in-a-power-of-a-sine-over-its-roots
Oct 16, 2026
Merged

Rafael-SOWNet merged 4 commits into
masterfrom
a-power-over-a-binomial-in-a-power-of-a-sine-over-its-roots

Conversation

@Rafael-SOWNet

Copy link
Copy Markdown
Member

Part of #718. Extends #1905.

#1905 splits 1/(a + b s^n) over the roots of the binomial, for s a sine, cosine or hyperbolic one. This takes a power of s above or below the bar as well:

  • k ≥ n: s^k/(a + b s^n) is s^(k - n)/b - (a/b) s^(k - n)/(a + b s^n).
  • k < 0: it is s^k/a - (b/a) s^(k + n)/(a + b s^n).
  • 0 < k < n: it is (1/(n a)) sum r^(k + 1)/(r - s) over all n roots.

So each term is the per-root closed form #1905 already writes, and the powers of s beside go to the integrator. It is asked early, where a power stands above or below: the rules before it spent the budget on sinh(y)^4/(a + b sinh(y)^3). The bare reciprocal is left to the reading #1905 gives it, and its answers are unchanged.

Fenced after the chain. At n = 4 with a power beside it, the rules before it answer sin(y)^3/(a - b sin(y)^4) in a tenth of a second, and asked first this took half of one. n = 4 is now theirs. The 33 rows the chain moved or slowed were re-run on both arms after the fence: the nine gains, and nothing slower.

DerivativeDiffersWhereReal decides a difference again in decimals. It counted a point where the answer's derivative had no value at the default precision as a difference. The derivative of the right answer to sinh(c + d x)^5/(a + b sinh(c + d x)^3) is NaN there at x = 3.17 with the pinned symbols, and the answer was refused. A difference, or no value, at the default precision is now decided again at forty and then eighty digits:

  • still no value where the integrand has one is still a difference;
  • a precision that cannot tell says nothing.

Checks

  • Exact checks at 50 and 300 digits on eleven shapes, three pin sets: 198/198.
  • AReciprocalOfABinomialInAPowerOfASineIntegralTest: 4 more cases.

Corpus, measured against master 90c9d3e (this branch then merged bf4be60):

master this branch
pocket: the algebraic pocket and every row of 4.1.7, 4.2.7, 6.1.7, 6.2.7 (8,745 statable) 8,696 8,705 before the fence
the 33 rows moved or slowed, re-run on both arms after the fence 23 32
rows gained +9: 6.1.7:200–204, 207–210

No losses. Gate passed; the suite on the fenced branch passed: 15,468, 0 failed.

🤖 Generated with Claude Code

https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura

Rafael-SOWNet and others added 4 commits October 16, 2026 14:27
…ower of it over its roots

s^k/(a + b s^n) for a whole k of either sign is divided down, or up, to 0 <= k < n over the
binomial, the powers of s beside going to the integrator, and is then
(1/(n a)) sum r^(k + 1)/(r - s) over all n roots. Asked first where a power stands above or
below, since the rules before it spent the budget: Rubi's 6.1.7 sinh(c + d x)^4/(a + b sinh(c + d x)^3),
csch(c + d x)^3/(a + b sinh(c + d x)^3) and the like timed out.

DerivativeDiffersWhereReal now decides a difference, or no value, found at the default precision
again in forty and eighty digits: the derivative of the right answer to
sinh(c + d x)^5/(a + b sinh(c + d x)^3) was NaN there at a sampled point, which counted as a
difference.

Part of #718

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura
They answer sin(y)^3/(a - b sin(y)^4) in a tenth of a second, where asked first the split took half of one.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura
…mial-in-a-power-of-a-sine-over-its-roots

# Conflicts:
#	Sources/AngouriMath/Functions/Continuous/Integration/Integration.Definition.cs
… changes

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura
@Rafael-SOWNet Rafael-SOWNet added this to the 2.6.0 milestone Oct 16, 2026
@Rafael-SOWNet
Rafael-SOWNet merged commit bc175c2 into master Oct 16, 2026
34 checks passed
@github-project-automation github-project-automation Bot moved this from Todo to Done in AngouriMath Oct 16, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Status: Done

Development

Successfully merging this pull request may close these issues.

1 participant