Poincaré-Lindstedt method

In perturbation theory, the Poincaré-Lindstedt method, named after Henri Poincaré, [H. Poincaré, Les Méthodes Nouvelles de la Mécanique Célèste I, II, III (Dover Publ., New York,1957).] and Anders Lindstedt [A. Lindstedt, Abh. K. Akad. Wiss. St. Petersburg 31, No. 4 (1882)] , is a technique for uniformly approximating periodic solutions to ordinary differential equations when regular perturbation approaches fail.

Example: the Duffing equation

The undamped, unforced Duffing equation is given by

:$ddot\{x\}\; +\; x\; +\; varepsilon\; x^3\; =\; 0,$

for $t>0$, with

:$x(0)\; =\; 1,\; dot\; x(0)\; =\; 0.,$

If we try to find an approximate solution of the form $x(t)=x\_0(t)\; +\; varepsilon\; x\_1(t)\; +\; cdots$, we obtain

:$x(t)\; =\; cos\; t\; +\; varepsilonleft(\; frac\{1\}\{32\}left(\; cos\; 3t\; -\; cos\; t\; ight)\; -\; frac\{3\}\{8\}t\; sin\; t\; ight).,$

This approximation grows without bound in time, which is inconsistent with the physical system that the equation models. The term responsible for this unbounded growth, called the "secular term", is $tsin\; t$. The Poincaré-Lindstedt method allows us to create an approximation that is accurate for all time, as follows.

In addition to expressing the solution itself as an asymptotic series, form another series with which to scale time $t$:

:$au\; =\; omega\; t,,$ where:$omega\; =\; 1\; +\; varepsilon\; omega\_1\; +\; cdots.,$

(Here we take $omega\_0\; =\; 1$ because the leading order of the solution's frequency is $1/2pi$.) Then the original problem becomes

:$omega^2\; x"(\; au)\; +\; x(\; au)\; +\; varepsilon\; x^3(\; au)\; =\; 0,$

with the same initial conditions. If we search for a solution of the form $x(\; au)=x\_0(\; au)\; +\; varepsilon\; x\_1(\; au)\; +\; cdots$, we obtain $x\_0\; =\; cos\; au$ and

:$x\_1\; =\; frac\{1\}\{32\}cos\; 3\; au\; +\; left(\; omega\_1\; -\; frac\{3\}\{8\}\; ight)\; ausin\; au.,$

So a secular term can be removed if we choose $omega\_1\; =\; 3/8$. We can continue in this way to higher orders of accuracy; as of now, we have the approximation

:$x(t)=cosleft(1\; +\; frac\{3\}\{8\}varepsilon\; ight)\; t\; +\; frac\{1\}\{32\}varepsiloncos\; 3left(1\; +\; frac\{3\}\{8\}varepsilon\; ight)t.\; ,$

References