Zulip Chat Archive

Stream: mathlib4

Topic: Define complex (path) integrals


Juanan (Nov 04 2023 at 20:00):

Hi,
I am trying to find a good definition for complex path integrals but can't come up with much. Do you know of any effective definition or have any suggestion?
Thanks!

Yury G. Kudryashov (Nov 04 2023 at 23:12):

What generality do you want?

Juanan (Nov 05 2023 at 11:43):

Yury G. Kudryashov said:

What generality do you want?

Hi Yury,
I am trying to define them for an arbitrary path γ and a given complex function. However, integration itself is not defined for complex numbers (except for contour integrals) so it's being quite tricky

Kalle Kytölä (Nov 05 2023 at 11:43):

Here is some earlier discussion on setting up complex contour integrals in a good generality. One of the main issues is that while it is easy to write down a definition for (piecewise) C1C^1 paths (and say continuous integrands), for integrals of holomorphic functions one really wants homotopy invariance in the usual C0C^0 sense, so the definition of contour integration had better make sense for C0C^0 paths and holomorphic integrands. So I think the ultimate mathlib-style definition is not yet in mathlib, but if you don't need homotopies and your paths are piecewise C1C^1 then it is not difficult to write down a direct definition.

Kalle Kytölä (Nov 05 2023 at 11:58):

For concreteness, something like the following would be ok for C1C^1 paths. When the mathlib-appropriate definition is made, there will certainly be a lemma showing that it coincides with this when γ\gamma is C1C^1 and ff is holomorphic (continuity of ff is "enough" but the general C0C^0 definition will only work for holomorphic integrands since it requires local exactness).

import Mathlib.Analysis.Complex.CauchyIntegral

/-- Contour integral of `f : ℂ → ℂ` along a path `γ : [a,b] → ℂ`. -/
noncomputable def contourIntegral (γ :   ) (a b : ) (f :   ) :  :=
   t in a..b, deriv γ t  f (γ t)

Juanan (Nov 05 2023 at 13:07):

Kalle Kytölä said:

For concreteness, something like the following would be ok for C1C^1 paths. When the mathlib-appropriate definition is made, there will certainly be a lemma showing that it coincides with this when γ\gamma is C1C^1 and ff is holomorphic (continuity of ff is "enough" but the general C0C^0 definition will only work for holomorphic integrands since it requires local exactness).

import Mathlib.Analysis.Complex.CauchyIntegral

/-- Contour integral of `f : ℂ → ℂ` along a path `γ : [a,b] → ℂ`. -/
noncomputable def contourIntegral (γ :   ) (a b : ) (f :   ) :  :=
   t in a..b, deriv γ t  f (γ t)

I'll try this!
Thank you so much Kalle :)


Last updated: Dec 20 2023 at 11:08 UTC