Skip to content

Integrate the incomplete gamma function of a whole order by parts against a reciprocal power - #2082

Merged
Rafael-SOWNet merged 1 commit into
masterfrom
differentiate-the-incomplete-gamma-function-by-parts
Oct 23, 2026
Merged

Rafael-SOWNet merged 1 commit into
masterfrom
differentiate-the-incomplete-gamma-function-by-parts

Conversation

@Rafael-SOWNet

Copy link
Copy Markdown
Member

Integrates gamma(n, a + b x) of a whole order n <= 0 over a power of another linear, by parts. Part of #2075.

The function is now a factor that integration by parts differentiates first, as the logarithms and the other special functions are: its derivative -u^(n - 1) e^(-u) u' is elementary. Against 1/(c + d x)^k with k >= 2, what is left is (a + b x)^(n - 1) e^(-a - b x) over a power of c + d x. That is a rational function times an exponential, which partial fractions and the exponential integral answer. One arm in IsDifferentiatedBeforeAPolynomial.

Rubi 8.6, 177 problems
master (aff8368d's integration rules, #2078) 144 solved
with this PR 156 solved, 0 wrong

Read with the harness copy that maps Maxima's GAMMA. The 12 rows are exactly the Gamma(n <= 0, a + b x)/(c + d x)^k ones, each answered in 0.15 to 0.32 s and 16 to 71 MB. No other row moved, and no integrand in the rest of the corpus holds the function.

The suite passes (16,894) and so do the MCP tests (329), measured before rebasing onto #2080, which touches nothing this reads. The new tests differentiate four of the rows back.

🤖 Generated with Claude Code

https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura

…inst a reciprocal power

Gamma(n, u) for u = a + b x is now a factor that integration by parts
differentiates first, as the logarithms and the other special functions are: its
derivative -u^(n - 1) e^(-u) u' is elementary. Against 1/(c + d x)^k with k >= 2
and a whole n <= 0, what is left is (a + b x)^(n - 1) e^(-a - b x) over a power of
c + d x, a rational function times an exponential, which partial fractions and
the exponential integral answer.

Rubi's 8.6, read with the harness copy that maps GAMMA: 144 -> 156 of 177, the
twelve Gamma(n <= 0, a + b x)/(c + d x)^k rows, 0 wrong, no other row moved, each
in 0.15 to 0.32 s. No integrand in the rest of the corpus holds the function.

Part of #2075.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura
@Rafael-SOWNet
Rafael-SOWNet merged commit e62a56d into master Oct 23, 2026
34 checks passed
@github-project-automation github-project-automation Bot moved this from Todo to Done in AngouriMath Oct 23, 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