Skip to content

Integrate a power of x times the logarithm of a polynomial plus a root, by parts - #1955

Merged
Rafael-SOWNet merged 1 commit into
masterfrom
by-parts-against-a-logarithm-of-a-radical-sum
Oct 20, 2026
Merged

Rafael-SOWNet merged 1 commit into
masterfrom
by-parts-against-a-logarithm-of-a-radical-sum

Conversation

@Rafael-SOWNet

Copy link
Copy Markdown
Member

Part of #233.

x^3 ln(-1 + 4 x + 4 sqrt((x - 1) x)) was declined, and three more of Rubi's 3.5 with half-odd powers of x beside the same logarithm ran out of time. Let D = A^2 - B^2 Q. The derivative of ln(A + B sqrt(Q)) is D'/(2 D) + (Q (A B' - A' B) + A B Q'/2)/(D sqrt(Q)), an identity of the algebra with no branch in it. By parts, what is left is a rational function plus one over the root, and the rules for those answer it. Here D = 8 x + 1.

The new rule is SolveByPartsAgainstALogarithmOfARadicalSum. It applies only for a D that is not constant. The inverse hyperbolic functions have a constant D, and they are answered already: x ln(x + sqrt(x^2 + 1)) is left to its existing route.

Measured against master 8a1ef454:

master this branch
Rubi family 3, all 1,382 rows (under machine load) 1,345 1,348
moves, each re-run alone 1 of 5 5 of 5, 0 wrong: 3.5:136, 143, 146 and 147. 3.2.2:152 was a load timeout that solves on both
families 0 (whole) and 1 (40 a file), samples of 2–8 identical
performance gate passed
suite 15,599 passed, 0 failed

The rebase onto #1950–#1954 conflicted only on the dispatch line, at the shared anchor. I kept both. The combined tests of #1952, #1953, #1954 and this branch pass, 17/17. All four rows were declined by 2.5.0, and BREAKING-CHANGES.md has the entry.

🤖 Generated with Claude Code

https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura

…t, by parts

With D = A^2 - B^2 Q, the derivative of ln(A + B sqrt(Q)) is
D'/(2 D) + (Q (A B' - A' B) + A B Q'/2)/(D sqrt(Q)), an identity of the
algebra with no branch in it, so by parts what is left is a rational
function and one over the root, which the rules for those answer. The search
differentiated the logarithm as written, a quotient with the root in both
parts, and ran past the budget. Only for a D that is not constant: the
inverse hyperbolic functions, whose D is, are answered already.

Rubi's 3.5: rows 136, 143, 146 and 147 go from unevaluated or timed out to
solved, 0 wrong. Families 0 and 1 and samples of 2-8 otherwise identical;
gate and suite pass.

Part of #233.

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 19, 2026
@Rafael-SOWNet
Rafael-SOWNet merged commit b53d799 into master Oct 20, 2026
34 checks passed
@github-project-automation github-project-automation Bot moved this from Todo to Done in AngouriMath Oct 20, 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