TOPICS
Search

Finite Part


The finite part of a divergent integral, also called the Hadamard finite part or partie finie, is the finite value obtained by removing the divergent terms from the asymptotic expansion of a specified cutoff integral and taking the remaining limit (Galapon 2016). The cutoff and subtraction prescription are part of the definition.

For a simple pole in the interior of the integration interval, the finite part obtained by symmetric exclusion of the pole agrees with the Cauchy principal value. The two notions differ in general. For example, symmetric exclusion about x=0 gives

 I(epsilon)=int_(-1)^(-epsilon)(dx)/(x^2)+int_epsilon^1(dx)/(x^2)=2/epsilon-2

for 0<epsilon<1. Since I(epsilon)->infty as epsilon->0^+, the Cauchy principal value does not exist as a finite number. Removing the divergent term instead gives

 FPint_(-1)^1(dx)/(x^2)=lim_(epsilon->0^+)[I(epsilon)-2/epsilon]=-2

(M. Onofri, pers. comm., Aug. 23, 2026).

The analogous prescription for partial sums gives a Hadamard-regularized sum.


See also

Cauchy Principal Value, Hadamard-Regularized Sum

Portions of this entry contributed by Mattia Onofri

Explore with Wolfram|Alpha

References

Galapon, E. A. "The Cauchy Principal Value and the Hadamard Finite-Part Integral as Values of Absolutely Convergent Integrals." J. Math. Phys. 57, 033502, 2016. https://doi.org/10.1063/1.4943300.

Referenced on Wolfram|Alpha

Finite Part

Cite this as:

Weisstein, Eric W., with contributions by Mattia Onofri. "Finite Part." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/FinitePart.html

Subject classifications