logo

What is the Laurent Series? 📂Complex Anaylsis

What is the Laurent Series?

Buildup

Taylor’s theorem generalizes the mean value theorem regarding the number of differentiations. It expands from dealing with something differentiated 11 times to nNn \in \mathbb{N} times. But if it was possible to generalize it to natural numbers, could it not be generalized to all integers? Of course, it’s not possible to differentiate n-n times, but what about considering integration, which is the inverse operation of differentiation? Here we introduce the Laurent’s Theorem without proof.

Assuming ff is analytic on two concentric circles C1:zα=r1\mathscr{C}_{1}: |z-\alpha| = r_{1} and C2:zα=r2\mathscr{C}_{2}: |z-\alpha| = r_{2} (r2<r1)(r_{2} < r_{1}) centered at the singularity α\alpha of f:ACCf: A \subset \mathbb{C} \to \mathbb{C}. Then, for all points between the two concentric circles, ff can be represented by f(z)=n=0an(zα)n+n=1bn(zα)n\displaystyle f(z) = \sum_{n = 0 }^{\infty} a_{n} (z-\alpha) ^{n} + \sum_{n = 1 }^{\infty} { {b_{n} } \over{ (z-\alpha) ^{n} } }.

  • an=12πiC1f(z)(zα)1+ndz,n=0,1,2,\displaystyle a_{n} = {{1} \over {2 \pi i}} \int_{\mathscr{C}_{1}} {{f(z)} \over {(z - \alpha)^{ 1 + n} }} dz \qquad , n = 0,1,2, \cdots
  • bn=12πiC2f(z)(zα)1ndz,n=1,2,3,\displaystyle b_{n} = {{1} \over {2 \pi i}} \int_{\mathscr{C}_{2}} {{f(z)} \over {(z - \alpha)^{ 1 - n} }} dz \qquad , n=1,2,3,\cdots

Definition

The following series is called the Laurent series. f(z)=n=0an(zα)n+n=1bn(zα)n f(z) = \sum_{n = 0 }^{\infty} a_{n} (z-\alpha) ^{n} + \sum_{n = 1 }^{\infty} { {b_{n} } \over{ (z-\alpha) ^{n} } }

Explanation

A Generalization of Cauchy’s Integral Formula for Differentiation: Let f:ACCf: A \subseteq \mathbb{C} \to \mathbb{C} be analytic in a simply connected region R\mathscr{R}.

If a simple closed path C\mathscr{C} in R\mathscr{R} encloses a point α\alpha, then for a natural number nn:

f(n)(α)n!=12πiCf(z)(zα)1+ndz {{f^{(n)} (\alpha) } \over {n!}} = {{1} \over {2 \pi i }} \int_{\mathscr{C}} {{f(z)} \over { (z - \alpha)^{1+n} }} dz

Using Cauchy’s Integral Formula will make it more evident that it is a generalization of Taylor’s theorem.

f(z)=n=0f(n)(α)n!(zα)n+n=1bn(zα)n f(z) = \sum_{n = 0 }^{\infty} {{f^{(n)} (\alpha) } \over {n!}} (z-\alpha) ^{n} + \sum_{n = 1 }^{\infty} { {b_{n} } \over{ (z-\alpha) ^{n} } } In such a series form, n=1bn(zα)n\displaystyle \sum_{n = 1 }^{\infty} { {b_{n} } \over{ (z-\alpha) ^{n} } } is referred to as the Principal Part. Notably, the coefficient of 1zα\displaystyle {{1} \over {z-\alpha}}, i.e., b1b_{1}, is defined as the Residue of ff at α\alpha and is expressed as b1=Resαf(z)b_{1} = \text{Res}_{\alpha} f(z)1.


  1. Osborne (1999). Complex variables and their applications: p144. ↩︎