PlanetMath (more info)
 Math for the people, by the people. Sponsor PlanetMath
Encyclopedia | Requests | Forums | Docs | Wiki | Random | RSS  
Login
create new user
name:
pass:
forget your password?
Main Menu
Owner confidence rating: Low Entry average rating: High
[parent] existence and uniqueness of solution to Cauchy problem (Theorem)

Let

\begin{displaymath}\begin{cases}\mathbf{\dot{x}} = F(\mathbf{x}, t) \\ \mathbf{x}(t_0) = \mathbf{x}_0\end{cases}\end{displaymath}
be a Cauchy problem, where $F : U \to \reals$ is

Then there exists a unique solution $\v x : I \to \reals^n$ of the Cauchy problem, defined in a neighborhood $I \subseteq \reals$ of $t_0$ .

Proof

Solving the Cauchy problem is equivalent to solving the following integral equation $$x(t) = x(t_0) + \int_{t_0}^{t} F(\v x(\tau), \tau) \d\tau$$

Let $X$ be the set of continuous functions $\v f : [t_0 - \delta, t_0 + \delta] \to B(\v x_0, \eps)$ . We'll assume $\eps$ to be chosen such that the $B(\v x_0, \eps) \subseteq U$ 1. In this ball, therefore, $F$ is Lipschitz continuous with respect to the first $n$ variable, in other words, there exists a real number $L$ such that $${F(\v x, t) - F(\v y, t)} \leq L\norm{\v x - \v y}$$ for all points $\v x, \v y$ sufficiently near to $\v x_0$ .

Now let's define the mapping $T : X \to X$ as follows $$T\v x : t \mapsto \v x_0 + \int_{t_0}^t F(\v x(\tau), \tau) \d \tau$$ We make the following observations about $T$ .

  1. Since $F$ is continuous, $\norm{F}$ attains a maximum value $M$ on the compact set $B(\v x_0, \eps) \times [t_0 \pm \delta]$ . But by hypothesis, $\norm{\v x(t) - \v x_0} \leq \eps$ , hence $$\norm{\v x(t) - \v x_0} \leq \int_{t_0}^t \norm{F(\v x(\tau), \tau)} \d \tau \leq M(t - t_0) \leq M\delta$$ for all $t \in [t_0 \pm \delta]$ .
  2. The Lipschitz continuity of $F$ yields $$\norm{T\v x(t) - T\v y(t)} \leq \int_{t_0}^t \norm{F(\v x(\tau), \tau) - F(\v y(\tau), \tau)} \d \tau \leq \int_{t_0}^t L\norm{\v x(\tau) - \v y(\tau)} \d \tau \leq L \delta d_\infty(\v x, \v y)$$
If we choose $\delta < \min\{1/L, \eps/M\}$ these conditions ensure that
  • $T(X) \subseteq X$ , i.e. $T$ doesn't send us outside of $X$ .
  • $T$ is a contraction mapping with respect to the uniform convergence metric $d_\infty$ on $X$ , i.e. there exists $\lambda \in \reals$ such that for all $\v x, \v y \in X$ , $$d_\infty(T\v x, T\v y) \leq \lambda d_\infty(\v x, \v x)$$
In particular, the second point allows us to apply Banach's theorem and define $$\v x^\star = \lim_{k \to \infty} T^k\v x_0$$ to find the unique fixed point of $T$ in $X$ , i.e. the unique function which solves $$T\v x = \v x \text{ in other words } \v x(t) = \v x_0 + \int_{t_0}^t F(\v x(\tau), \tau) \d \tau$$ and which therefore locally solves the Cauchy problem.



Footnotes

... $B(\v x_0, \eps) \subseteq U$1
$B(\v x_0, \eps)$ denotes the closed ball $\{ \v x : \norm{\v x_0 - \v x} \leq \eps \}$



"existence and uniqueness of solution to Cauchy problem" is owned by ehremo.
(view preamble | get metadata)

View style:

Keywords:  cauchy initial value problem existence uniqueness solution

This object's parent.
Log in to rate this entry.
(view current ratings)

Cross-references: function, fixed point, theorem, metric, uniform convergence, contraction mapping, hypothesis, compact set, mapping, near, points, real number, ball, closed ball, integral equation, equivalent, proof, solution, Lipschitz continuous, neighborhood, variables, continuous function, Cauchy problem

This is version 17 of existence and uniqueness of solution to Cauchy problem, born on 2007-04-10, modified 2007-08-12.
Object id is 9173, canonical name is ExistenceAndUniquenessOfSolutionToCauchyProblem.
Accessed 1266 times total.

Classification:
AMS MSC34A12 (Ordinary differential equations :: General theory :: Initial value problems, existence, uniqueness, continuous dependence and continuation of solutions)

Pending Errata and Addenda
None.
[ View all 1 ]
Discussion
Style: Expand: Order:
forum policy

No messages.

Interact
post | correct | update request | prove | add result | add corollary | add example | add (any)