The analysis of Partial Differential Equations (PDEs) relies heavily on fixed point strategies to transition from linear solvability to nonlinear theory. While linear equations often possess explicit solutions (via Green's [functions](/page/Function) or [Fourier transforms](/page/Fourier%20Transform)), nonlinear problems require an iterative or [topological](/page/Topology) approach. The fundamental idea is to rewrite a PDE of the form $L(u) = N(u)$ as a fixed point problem $u = T(u)$, where $T$ is an operator acting on a functional space $X$. The existence of a solution corresponds to finding $\overline{x} \in X$ such that $T(\overline{x}) = \overline{x}$.

We examine three primary classes of fixed point theorems:

1. **Contraction Mapping:** Guarantees existence and uniqueness, typically used for local-in-time solutions. 2. **Compactness Methods:** Guarantees existence without uniqueness, used when contraction fails but a priori bounds exist. 3. **Variational Inequalities:** Handles problems with constraints using monotonicity.

---

## 1. The [Contraction Mapping Principle](/page/Contraction%20Mapping%20Principle)

This method is the cornerstone of local existence theory, relying on the metric completeness of the space.

[definition:Contraction] Let $(X, d)$ be a [metric space](/page/Metric%20Space). A map $f: X \to X$ is called a **contraction** if there exists $\lambda < 1$ such that: \begin{align*} d(f(x), f(y)) \le \lambda d(x, y) \quad \forall x, y \in X. \end{align*} [/definition]

[theorem:Banach Contraction Principle] Let $f$ be a contraction on a complete metric space $X$. Then $f$ has a unique fixed point $\overline{x} \in X$. [/theorem]

### Application: Semilinear Parabolic Equations

We apply this to a semilinear [heat equation](/page/Heat%20Equation).

[example:Parabolic Cauchy Problem] **Problem:** Let $\Omega \subset \mathbb{R}^3$ be a bounded domain with smooth [boundary](/page/Boundary). We seek a solution $u$ to the problem: \begin{align*} \begin{cases} \partial_t u - \Delta u = h(u) & \text{in } (0, t_0] \times \Omega \\ u = 0 & \text{on } (0, t_0] \times \partial \Omega \\ u(0, x) = u_0(x) & \text{in } \Omega \end{cases} \end{align*} where $h(u) = \pm |u|^{p-1}u$ with $p \in [1, 5)$ and $u_0 \in H^1_0(\Omega)$. **Solution:** We define the solution space $X_{t_0} = C([0, t_0]; H^1_0(\Omega)) \cap L^2(0, t_0; H^2(\Omega))$. We define a map $f: B \to X_{t_0}$ on a closed ball $B$ of radius $4R$, where $u = f(w)$ is the unique solution to the linear problem $\partial_t u - \Delta u = h(w)$. To prove $f$ is a contraction, we analyze the difference $\overline{u} = u_1 - u_2$ for inputs $w_1, w_2$. Using energy estimates and Agmon's inequality (valid for $p < 5$), we obtain: \begin{align*} \|\overline{u}\|_{X_{t_0}} \le \lambda \|\overline{w}\|_{X_{t_0}} \end{align*} where $\lambda$ depends on $t_0^{1/(5-p)}$. By choosing $t_0$ sufficiently small, we ensure $\lambda < 1$, guaranteeing a unique strong solution. [/example]

---

## 2. Fixed Point Theorems Without Uniqueness

In infinite-dimensional spaces, closed balls are not compact, preventing the direct use of Brouwer's theorem. We instead require the operator itself to be compact.

[theorem:Schauder Fixed Point Theorem] Let $X$ be a [Banach space](/page/Banach%20Space), and let $C \subset X$ be a nonempty, closed, bounded, and convex set. If $f: C \to C$ is continuous and compact (i.e., $f(C)$ is relatively compact in $X$), then $f$ has a fixed point. [/theorem]

A powerful generalization that incorporates *a priori* bounds is Schaefer's theorem.

[theorem:Schaefer Fixed Point Theorem] Let $X$ be a Banach space and let $f: X \to X$ be a compact operator. Suppose that the [set](/page/Set) of possible fixed points for the homotopy: \begin{align*} \mathcal{F} := \{ x \in X : x = \lambda f(x) \text{ for some } \lambda \in [0, 1] \} \end{align*} is bounded. Then $f$ has a fixed point in $X$. [/theorem]

### Application: Semilinear Elliptic Equations

[example:Elliptic Dirichlet Problem] **Problem:** Solve the nonlinear elliptic problem: \begin{align*} \begin{cases} -\Delta u + |u|^{p-1}u = g & \text{in } \Omega \\ u = 0 & \text{on } \partial \Omega \end{cases} \end{align*} where $g \in H^{-1}(\Omega)$ and $p \in [1, 5)$. **Solution:** We define the map $f: H^1_0(\Omega) \to H^1_0(\Omega)$ such that $u = f(w)$ solves $-\Delta u = g - h(w)$. 1. **Compactness:** The map $w \mapsto h(w)$ maps bounded sets in $H^1_0$ to bounded sets in $L^{(p+1)/p}$, which compactly embed into $H^{-1}$. Thus $f$ is a compact operator. 2. **A Priori Bounds:** We consider the equation $u_\lambda = \lambda f(u_\lambda)$, which implies: \begin{align*} -\Delta u_\lambda = \lambda g - \lambda h(u_\lambda) \end{align*} Multiplying by $u_\lambda$ and integrating yields $\|\nabla u_\lambda\|_{L^2}^2 \le \lambda \langle g, u_\lambda \rangle$, implying $\|u_\lambda\|_{H^1} \le \|g\|_{H^{-1}}$. Since the set $\mathcal{F}$ is bounded, a solution exists. [/example]

---