Symbolic Integration in Weierstrass-like Extensions

Technical Summary: Symbolic Integration in Weierstrass-like Extensions

Problem Statement
The paper addresses the problem of symbolic integration within differential fields containing functions defined by first-order nonlinear differential equations. Specifically, it focuses on "Weierstrass-like extensions," which generalize the classical Weierstrass $\wp$-function. These extensions are generated by an element $t$ satisfying an equation of the form $(t')^2 = \alpha^2(4t^3 - g_2t - g_3)$ (or more generally, where $t'$ is integral over $k[t]$ and satisfies a minimal polynomial $m(t, X)$). While algorithms exist for elementary functions and certain algebraic functions, the integration of powers of $\wp$ and similar functions has lacked a comprehensive algorithmic framework, with few known identities for such integrals in the literature.

Methodology
The authors employ a differential algebraic approach, extending the theory of symbolic integration from rational and elementary functions to these specific algebraic extensions. The methodology proceeds through three distinct reduction stages, analogous to the Risch algorithm's structure but adapted for the non-trivial algebraic nature of the extensions:

1. Algebraic and Differential Foundations: The paper establishes the necessary terminology regarding places, valuation rings, and ramification indices in the field $K = k(t, t')$. A critical innovation is the refinement of the concept of "special polynomials" (polynomials whose roots are "special points" where the derivative vanishes in the residue field) to this context. The authors distinguish between "normal" places and "special places," further categorizing special places into "zeroth kind" (where the order of the derivative is strictly less than the ramification index) and first kind. This distinction is vital for controlling the orders of poles after differentiation.

2. Hermite Reduction: The first algorithmic step decomposes the integrand into a normal part and a special part. The normal part (denominators composed of normal irreducible factors) is reduced using a generalized Hermite reduction. This process decreases the multiplicity of normal factors in the denominator modulo derivatives. The authors prove that this reduction yields a "Hermite remainder" with a squarefree normal denominator.

3. Special Reduction: The special part (denominators composed of special factors) is handled under Hypothesis 16, which assumes that all special points of the minimal polynomial are constants in the base field $k$. Under this hypothesis, the authors develop a reduction algorithm that decreases the multiplicity of special factors in the denominator. This involves solving linear congruences to eliminate higher-order poles, resulting in a "special remainder" of the form $\frac{\theta}{\gamma}t'$, where $\gamma$ is squarefree and composed only of special factors.

4. Polynomial Reduction: The remaining terms, which are integral over $k[t]$ and have no special parts, are processed via polynomial reduction. This step controls the behavior at infinity by reducing the degree of the polynomial part of the integrand. The authors define a linear operator $S_*$ and iteratively subtract derivatives of specific forms to lower the degree of the polynomial remainder until it falls below a threshold determined by the degree of the defining polynomial $q$.

Key Contributions

• Extension of Special Polynomials: The paper generalizes Bronstein's notion of special polynomials to Weierstrass-like extensions, introducing the concept of "special places of the zeroth kind," which are unique to these non-monomial algebraic extensions.
• Algorithmic Decomposition: The authors present a complete decomposition theorem (Theorem 25). For any element $f$ in the extension $K$, they prove the existence of a decomposition:
  $$f = g' + h + s + l + \eta$$
  where $g \in K$, $h$ is the Hermite remainder, $s$ is the special remainder, $l$ is a unique element in the integral basis with no special part, and $\eta$ is a polynomial remainder.
• Integrability Criteria: The decomposition provides a necessary and sufficient condition for in-field integrability: $f$ has an elementary integral in $K$ if and only if the remainders $h, s, l$ are zero and $\eta$ is the derivative of an element in the base field $k$.
• Elementary Integrability Bound: The paper derives a necessary condition for the existence of an elementary integral over $K$: the degree of the polynomial remainder $\eta$ must satisfy $\deg_t \eta \leq \frac{\deg_t q}{2} - 1$.

Results
Applying these algorithms to the powers of the Weierstrass $\wp$-function (where $t = \wp(z)$), the authors derive new recurrence relations and explicit formulas for integrals of the form $\int \wp(z)^n dz$.

• They confirm that $\int \wp(z) dz$ is not elementary over $C(z)(\wp, \wp')$, necessitating the introduction of the Weierstrass zeta function $\zeta(z)$.
• They provide explicit closed-form expressions for $\int \wp(z)^2 dz$, $\int \wp(z)^3 dz$, and $\int \wp(z)^4 dz$, expressing them in terms of $\wp, \wp', \zeta, z$, and the invariants $g_2, g_3$.
• They generalize these results to an arbitrary power $n$, showing how the recurrence relations for the integrals can be systematically derived using the proposed reduction algorithms.

Significance
The paper claims significance primarily in the theoretical and algorithmic unification of integration for a class of functions defined by nonlinear first-order differential equations. By extending the classical notion of special polynomials and providing a rigorous reduction framework, the work fills a gap in the literature regarding the integration of $\wp$-function powers. The authors position their results as a continuation of the trend in developing integration algorithms for increasingly complex function classes. The derived formulas for powers of $\wp$ serve as a concrete application, demonstrating the utility of the proposed reductions in generating new identities for integrals that were previously not systematically computable within this algebraic framework.