Green's function: Difference between revisions
imported>Paul Wormer (saved without checking LaTeX: processor down) |
imported>Paul Wormer (saved without checking LaTeX: processor down on and off) |
||
Line 1: | Line 1: | ||
In [[physics]] and [[mathematics]], '''Green's | In [[physics]] and [[mathematics]], '''Green's functions''' are auxiliary functions in the solution of linear partial [[differential equations]]. Green's function is named for the British mathematician [[George Green]] (1793 – 1841). | ||
==Definition== | |||
Let ''L''<sub>'''''x'''''</sub> be a given linear differential operator in ''n'' variables '''''x''''' = (''x''<sub>1</sub>, ''x''<sub>2</sub>, ..., ''x''<sub>''n''</sub>), then the ''Green function of'' ''L''<sub>'''''x'''''</sub> is the function ''G''('''''x''''','''''y''''') defined by | Let ''L''<sub>'''''x'''''</sub> be a given linear differential operator in ''n'' variables '''''x''''' = (''x''<sub>1</sub>, ''x''<sub>2</sub>, ..., ''x''<sub>''n''</sub>), then the ''Green function of'' ''L''<sub>'''''x'''''</sub> is the function ''G''('''''x''''','''''y''''') defined by | ||
:<math> | :<math> | ||
L_\mathbf{x} G(\mathbf{x},\mathbf{y}) = | L_\mathbf{x} G(\mathbf{x},\mathbf{y}) = \delta(\mathbf{x}- \mathbf{y}), | ||
</math> | </math> | ||
where δ('''''x'''''-'''''y''''') is the [[Dirac delta function]]. Once ''G''('''''x''''','''''y''''') is known, any differential equation involving ''L''<sub>'''''x'''''</sub> is formally solved, | where δ('''''x'''''-'''''y''''') is the [[Dirac delta function]]. Once ''G''('''''x''''','''''y''''') is known, any differential equation involving ''L''<sub>'''''x'''''</sub> is formally solved. Suppose we want to solve, | ||
:<math> | |||
L_\mathbf{x} \,\phi(\mathbf{x}) = \rho(\mathbf{x}) | |||
</math> | |||
for a known right hand side ρ('''''x'''''). | |||
The formal solution is | |||
:<math> | :<math> | ||
\phi(\mathbf{x}) = \int\; G(\mathbf{x},\mathbf{y})\; \rho(\mathbf{y})\; \mathrm{d}\mathbf{y}. | \phi(\mathbf{x}) = \int\; G(\mathbf{x},\mathbf{y})\; \rho(\mathbf{y})\; \mathrm{d}\mathbf{y}. | ||
</math> | </math> | ||
The proof is by verification, | The proof is by verification, | ||
:<math> | :<math> | ||
L_\mathbf{x} \,\phi(\mathbf{x}) = \int\;L_\mathbf{x} \; G(\mathbf{x},\mathbf{y})\; \rho(\mathbf{y})\; \mathrm{d}\mathbf{y} = | L_\mathbf{x} \,\phi(\mathbf{x}) = \int\;L_\mathbf{x} \; G(\mathbf{x},\mathbf{y})\; \rho(\mathbf{y})\; \mathrm{d}\mathbf{y} = \int\;\delta(\mathbf{x}- \mathbf{y})\;\rho(\mathbf{y}) \mathrm{d}\mathbf{y} = \rho(\mathbf{x}) | ||
</math> | </math> | ||
where in the last step the defining property of the Dirac delta function is used. | where in the last step the defining property of the Dirac delta function is used. | ||
The Green function | The integral operator that has the Green function as kernel may be seen as the inverse of a linear operator, | ||
:<math> | :<math> | ||
\ | L_\mathbf{x}\;\phi(\mathbf{x}) = \rho(\mathbf{x}) \quad\Longrightarrow \quad \phi( \mathbf{x}) =L_\mathbf{x}^{-1}\;\rho(\mathbf{x}) = \int G(\mathbf{x},\mathbf{y}) \rho(\mathbf{y})\;\mathrm{d}\mathbf{y} . | ||
</math> | </math> | ||
It is illuminating to make the analogy with matrix equations. Let <math>\mathbb{L}</math> and <math>\mathbb{G}</math> be ''n''×''n'' matrices connected by | |||
:<math> | :<math> | ||
\mathbb{L}\boldsymbol{\phi} = | \mathbb{L} \mathbb{G} = \mathbb{E}\quad \Longleftrightarrow \quad \left(\mathbb{L} \mathbb{G}\right)_{ij} = \delta_{ij}, \quad\hbox{i.e.,}\quad \mathbb{G} = \mathbb{L}^{-1}, | ||
\ | </math> | ||
then the solution of a matrix-vector equation is | |||
:<math> | |||
\mathbb{L}\boldsymbol{\phi} = \boldsymbol{\rho}\quad \Longrightarrow \quad | |||
\phi_i = \sum_{j} \mathbb{G}_{ij} \rho_j. | |||
</math> | </math> | ||
Make the correspondence ''i'' ↔ '''''x''''', ''j'' ↔ '''''y''''', and compare the sum over ''j'' with the integral over '''''y''''', and the correspondence is evident. | |||
==Example== | |||
We consider a case of three variables, ''n'' = 3. The Green function of | |||
:<math> | |||
-\frac{1}{4\pi} \boldsymbol{\nabla}^2 \equiv -\frac{1}{4\pi} | |||
\left( \frac{\partial^2}{\partial x^2}+\frac{\partial^2}{\partial y^2}+\frac{\partial^2}{\partial z^2}\right) | |||
</math> | |||
is | |||
:<math> | |||
G(\mathbf{x},\mathbf{y}) = \frac{1}{|\mathbf{x}-\mathbf{y}|}. | |||
</math> | |||
'''(To be continued)''' | '''(To be continued)''' | ||
==Reference== | ==Reference== | ||
P. Roman, ''Advanced Quantum Theory'', Addison-Wesley, Reading, Mass. (1965) Appendix 4. | P. Roman, ''Advanced Quantum Theory'', Addison-Wesley, Reading, Mass. (1965) Appendix 4. |
Revision as of 08:35, 8 January 2009
In physics and mathematics, Green's functions are auxiliary functions in the solution of linear partial differential equations. Green's function is named for the British mathematician George Green (1793 – 1841).
Definition
Let Lx be a given linear differential operator in n variables x = (x1, x2, ..., xn), then the Green function of Lx is the function G(x,y) defined by
where δ(x-y) is the Dirac delta function. Once G(x,y) is known, any differential equation involving Lx is formally solved. Suppose we want to solve,
for a known right hand side ρ(x). The formal solution is
The proof is by verification,
where in the last step the defining property of the Dirac delta function is used.
The integral operator that has the Green function as kernel may be seen as the inverse of a linear operator,
It is illuminating to make the analogy with matrix equations. Let and be n×n matrices connected by
then the solution of a matrix-vector equation is
Make the correspondence i ↔ x, j ↔ y, and compare the sum over j with the integral over y, and the correspondence is evident.
Example
We consider a case of three variables, n = 3. The Green function of
is
(To be continued)
Reference
P. Roman, Advanced Quantum Theory, Addison-Wesley, Reading, Mass. (1965) Appendix 4.