# Integrability in the complex case

Line 6: | Line 6: | ||

Therefore, we have to define <math>curves</math> and <math>paths</math> in <math>\mathbb {C}</math>. | Therefore, we have to define <math>curves</math> and <math>paths</math> in <math>\mathbb {C}</math>. | ||

− | + | {{Definition|A ''curve'' in <math>\mathbb {C}</math> is a function <math>\gamma :[a,b]\rightarrow \mathbb {C}</math> such that: | |

− | <math display="block"> \forall \, t\in [a,b],\; \gamma (t)\in \mathbb {C}. </math | + | <math display="block"> \forall \, t\in [a,b],\; \gamma (t)\in \mathbb {C}. </math> |

− | + | }} | |

− | < | + | {{Definition| |

+ | The set <math>\gamma ^{\star }=\gamma ([a,b])=\{ \gamma (t)\in \mathbb {C}\colon t\in [a,b]\} </math> is edfined as the ''support of the curve'' <math>\gamma </math>. The points <math>\gamma (a)</math> and <math>\gamma (b)</math> are respectivelty the first and the second extremum of the curve. If <math>\gamma (a)=\gamma (b)</math> we say that the curve is ''closed.'' | ||

+ | }} | ||

+ | |||

+ | {{Definition| A curve <math>\gamma </math> in <math>\mathbb {C}</math> is a ''path'' if it is piecewise differentiable. | ||

+ | }} | ||

Now, we can give a meaning to the expression ''integrate along a path''. The definition we will give is the same given in the course of Multivariable Calculus for the elements of an oriented path. All the remarks given in the case of a multivariable function follow: among them, we have the invariance of the integral with respect to the chosen oriented path. In the following, we can take <math>[a,b]\equiv [0,1]</math>. | Now, we can give a meaning to the expression ''integrate along a path''. The definition we will give is the same given in the course of Multivariable Calculus for the elements of an oriented path. All the remarks given in the case of a multivariable function follow: among them, we have the invariance of the integral with respect to the chosen oriented path. In the following, we can take <math>[a,b]\equiv [0,1]</math>. | ||

− | + | {{Definition|Let <math>\gamma :[a,b]\rightarrow \mathbb {C}</math> be a path and <math>f:\mathcal{D}\subset \mathbb {C}\rightarrow \mathbb {C}</math> a complex-valued function. We define ''integral of f along the path'' <math>\gamma </math> the quantity: | |

− | <math display="block"> \int _{\gamma }f(z)dz=\int _{a}^{b}f(\gamma (t))\gamma ^{\prime }(t)dt. </math> | + | <math display="block"> \int _{\gamma }f(z)dz=\int _{a}^{b}f(\gamma (t))\gamma ^{\prime }(t)dt. </math> |

+ | }} | ||

− | + | {{Definition| | |

+ | We define the length of the curve ( or the path ) <math>\gamma </math> as the quantity: | ||

− | <math display="block"> L(\gamma )\equiv \int _{0}^{1}\mid \gamma ^{\prime }(t)\mid dt. </math> | + | <math display="block"> L(\gamma )\equiv \int _{0}^{1}\mid \gamma ^{\prime }(t)\mid dt. </math> |

+ | }} | ||

We have that: | We have that: | ||

Line 29: | Line 37: | ||

An important concept, that distinguishes the integral along paths in <math>\mathbb {C}</math> from the one in <math>\mathbb {R}^{2}</math> is the ''index.'' In particular, we have the following: | An important concept, that distinguishes the integral along paths in <math>\mathbb {C}</math> from the one in <math>\mathbb {R}^{2}</math> is the ''index.'' In particular, we have the following: | ||

− | + | {{Definition| | |

+ | Let <math>\gamma :[a,b]\rightarrow \mathbb {C}</math> be a closed path and <math>\Omega =\mathbb {C}\setminus \gamma ^{\star }</math>. <math>\forall \, z\in \Omega </math> we define ''index of z with respect to'' <math>\gamma </math> the quantity: | ||

− | <math display="block"> Ind_{\gamma }(z)=\frac{1}{2\pi i}\oint _{\gamma }\frac{d\xi }{\xi -z}\equiv \frac{1}{2\pi i}\oint dt\frac{\gamma ^{\prime }(t)}{\gamma (t)-z}. </math> | + | <math display="block"> Ind_{\gamma }(z)=\frac{1}{2\pi i}\oint _{\gamma }\frac{d\xi }{\xi -z}\equiv \frac{1}{2\pi i}\oint dt\frac{\gamma ^{\prime }(t)}{\gamma (t)-z}. </math> |

+ | }} | ||

In this defintion we go through <math>\gamma </math> in a counter-clockwise way. We set <math>Ind_{\gamma }(\infty )=0.</math> | In this defintion we go through <math>\gamma </math> in a counter-clockwise way. We set <math>Ind_{\gamma }(\infty )=0.</math> | ||

Line 37: | Line 47: | ||

There is an important relation between the function <math>Ind_{\gamma }(z)</math> and the condition of holomorphism, as given by the following | There is an important relation between the function <math>Ind_{\gamma }(z)</math> and the condition of holomorphism, as given by the following | ||

− | + | {{Theorem| | |

+ | <math>Ind_{\gamma }(z)</math> is an holomorphic function in <math>\mathbb {C}</math>. Moreover, we have that <math>Ind_{\gamma }\colon \Omega \subset \mathbb {C}\rightarrow \mathbb {Z}</math>, that is <math>Ind_{\gamma }</math> is an interger valued function and has constant valued on each connencted component in which <math>\mathbb {C}</math> is splitted by <math>\gamma </math>. On the exterior of <math>\gamma </math> the index is always zero. | ||

+ | }} | ||

As for the definition of interior and exterior, we follow the ''Jordan Theorem'' for simple and closed curves in <math>\mathbb {R}^{2}.</math> We can refer to books of Multivariate Calucus about this topic. | As for the definition of interior and exterior, we follow the ''Jordan Theorem'' for simple and closed curves in <math>\mathbb {R}^{2}.</math> We can refer to books of Multivariate Calucus about this topic. | ||

Line 43: | Line 55: | ||

We provide a proof of the previous theorem in a simple case: | We provide a proof of the previous theorem in a simple case: | ||

− | + | {{Proof|Let <math>\xi \in \mathbb {C}</math> and <math>\gamma </math> be a circumference with centre <math>z_{0}</math> and radius <math>r</math>. So, we have that <math>\gamma (t)=re^{it}+z_{0},\, t\in [0,2\pi ].</math> | |

We obtain that: | We obtain that: | ||

<math display="block"> Ind_{\gamma }(z)=\frac{1}{2\pi i}\oint _{\gamma }\frac{d\xi }{\xi -z}=\frac{1}{2\pi i}\int _{0}^{2\pi }\frac{\gamma ^{\prime }(t)}{\gamma (t)-z}dt=\frac{1}{2\pi i}\int _{0}^{2\pi }\frac{ire^{it}}{re^{it}}dt=1. </math> | <math display="block"> Ind_{\gamma }(z)=\frac{1}{2\pi i}\oint _{\gamma }\frac{d\xi }{\xi -z}=\frac{1}{2\pi i}\int _{0}^{2\pi }\frac{\gamma ^{\prime }(t)}{\gamma (t)-z}dt=\frac{1}{2\pi i}\int _{0}^{2\pi }\frac{ire^{it}}{re^{it}}dt=1. </math> | ||

− | + | }} | |

We remind the defintions of ''connected, convex & simply connected'' set. | We remind the defintions of ''connected, convex & simply connected'' set. | ||

− | + | {{Definition|We say that <math>\Omega \subset \mathbb {C}</math> is connected if:<br/> <math>\forall \, z_{1},z_{2}\in \Omega </math> we have that <math>z(t)=(1-t)z_{1}+tz_{2}\subset \Omega .</math> | |

+ | }} | ||

In general, for the other twoue definitions, we refer to textbooks of Multivariate Calculus, where one can find rigorous definitions for <math>\Omega \subset \mathbb {R}^{n},\, n\geq 2</math> that can be easily extended to <math>\mathbb {C}.</math> | In general, for the other twoue definitions, we refer to textbooks of Multivariate Calculus, where one can find rigorous definitions for <math>\Omega \subset \mathbb {R}^{n},\, n\geq 2</math> that can be easily extended to <math>\mathbb {C}.</math> | ||

Line 60: | Line 73: | ||

<math display="block"> \hat{\mathbb {C}}\equiv \mathbb {C}\cup \{ \infty \} . </math> | <math display="block"> \hat{\mathbb {C}}\equiv \mathbb {C}\cup \{ \infty \} . </math> | ||

− | + | {{Theorem| We consider <math>\Omega \subset \hat{\mathbb {C}}</math> and <math>f</math> an holomorphic function in <math>\Omega .</math> If <math>\gamma </math> is a closed path in <math>\Omega </math> such that <math>Ind_{\gamma }(z)=0,\, \forall \, z\in \hat{\mathbb {C}}\setminus \Omega ,</math> then: | |

<math display="block"> \int _{\gamma }f(z)dz=0. </math> | <math display="block"> \int _{\gamma }f(z)dz=0. </math> | ||

Line 68: | Line 81: | ||

Let <math>\Omega \subset \hat{\mathbb {C}}</math> be a ''simply connected'' set (and so <math>\hat{\mathbb {C}}\setminus \Omega </math> is connected) and <math>f</math> be a holomorphic function in <math>\Omega .</math> If <math>\gamma </math> is a closed path in <math>\Omega </math>, then: | Let <math>\Omega \subset \hat{\mathbb {C}}</math> be a ''simply connected'' set (and so <math>\hat{\mathbb {C}}\setminus \Omega </math> is connected) and <math>f</math> be a holomorphic function in <math>\Omega .</math> If <math>\gamma </math> is a closed path in <math>\Omega </math>, then: | ||

− | <math display="block"> \int _{\gamma }f(z)dz=0. </math> | + | <math display="block"> \int _{\gamma }f(z)dz=0. </math> |

+ | }} | ||

We see that the theorem provides a ''necessary condition'' for the integral along the path to be zero, in fact we could have cases in which the curve has a non-zero index and so we could not say anything about the value of the integral a priori. | We see that the theorem provides a ''necessary condition'' for the integral along the path to be zero, in fact we could have cases in which the curve has a non-zero index and so we could not say anything about the value of the integral a priori. | ||

Line 74: | Line 88: | ||

An important corollary of Cauchy theorem gives conditions for the comparison between integrals of the same holomorphic <math>f</math> along two different closed paths in <math>\Omega </math>. | An important corollary of Cauchy theorem gives conditions for the comparison between integrals of the same holomorphic <math>f</math> along two different closed paths in <math>\Omega </math>. | ||

− | + | {{Theorem|We consider <math>\Omega \subset \hat{\mathbb {C}}</math> and <math>f</math> an holomorphic function in <math>\Omega .</math> Let <math>\gamma _{1}\; \& \; \gamma _{2}</math> be two closed paths in <math>\Omega </math> such that <math>Ind_{\gamma _{1}}(z)=Ind_{\gamma _{2}}(z),\; \forall \, z\in \hat{\mathbb {C}}\setminus \Omega </math>, then: <math display="block"> \int _{\gamma _{1}}f(z)dz=\int _{\gamma _{2}}f(z)dz. </math> | |

+ | }} | ||

− | + | {{Proof| Let <math>f</math> be holomorphic in <math>\Omega </math>and <math>\gamma ,\lambda </math> be two closed paths. | |

We link <math>\gamma </math> and <math>\lambda </math> with two line segments <math>\pm c</math> with a distance <math>\epsilon </math> between them. | We link <math>\gamma </math> and <math>\lambda </math> with two line segments <math>\pm c</math> with a distance <math>\epsilon </math> between them. | ||

Line 91: | Line 106: | ||

It follows that: <math>\int _{\gamma }f(z)dz=\int _{\lambda }f(z)dz.</math> | It follows that: <math>\int _{\gamma }f(z)dz=\int _{\lambda }f(z)dz.</math> | ||

− | + | }} |

## Revision as of 08:51, 21 October 2016

We introduce the theory of integrable functions in . From the integrability of a function it will follow that it is differentiable infinite times, where it is holomorphic.

We will not consider generic integrals, but *integrals along paths.*

Therefore, we have to define and in .

**Definition**

A *curve* in is a function such that:

**Definition**

*support of the curve*. The points and are respectivelty the first and the second extremum of the curve. If we say that the curve is

*closed.''*

**Definition**

*path*if it is piecewise differentiable.

Now, we can give a meaning to the expression *integrate along a path*. The definition we will give is the same given in the course of Multivariable Calculus for the elements of an oriented path. All the remarks given in the case of a multivariable function follow: among them, we have the invariance of the integral with respect to the chosen oriented path. In the following, we can take .

**Definition**

Let be a path and a complex-valued function. We define *integral of f along the path* the quantity:

**Definition**

We define the length of the curve ( or the path ) as the quantity:

We have that:

An important concept, that distinguishes the integral along paths in from the one in is the *index.* In particular, we have the following:

**Definition**

Let be a closed path and . we define *index of z with respect to* the quantity:

In this defintion we go through in a counter-clockwise way. We set
We can see that every paths in satisfy an important property, analogous to the one given by the *Jordan Theorem* for simple and closed curves in .In fact, we have that always divides in two connected components, one bounded and the other unbounded.
There is an important relation between the function and the condition of holomorphism, as given by the following

**Theorem**

As for the definition of interior and exterior, we follow the *Jordan Theorem* for simple and closed curves in We can refer to books of Multivariate Calucus about this topic.

We provide a proof of the previous theorem in a simple case:

**Proof:** * *

Let and be a circumference with centre and radius . So, we have that We obtain that:

We remind the defintions of *connected, convex & simply connected* set.

**Definition**

we have that

In general, for the other twoue definitions, we refer to textbooks of Multivariate Calculus, where one can find rigorous definitions for that can be easily extended to

We state two of the most important results in complex analysis, given the notion:

**Theorem**

We consider and an holomorphic function in If is a closed path in such that then:

Equivalently, we have that:

Let be a *simply connected* set (and so is connected) and be a holomorphic function in If is a closed path in , then:

We see that the theorem provides a *necessary condition* for the integral along the path to be zero, in fact we could have cases in which the curve has a non-zero index and so we could not say anything about the value of the integral a priori.

An important corollary of Cauchy theorem gives conditions for the comparison between integrals of the same holomorphic along two different closed paths in .

**Theorem**

**Proof:** * *

Let be holomorphic in and be two closed paths. We link and with two line segments with a distance between them. It is not possible to have a closed path with zero index with respect to every . Let such given by: By applying Cauchy's theorem to the function considered: