Open Mapping and Closed Graph Theorems - Key Properties
The Closed Graph Theorem provides a powerful criterion for boundedness of linear operators, transforming a sequential condition into a norm estimate.
Let and be Banach spaces and a linear operator with closed graph. Then is bounded.
This theorem is extremely useful in practice because verifying that a graph is closed (a sequential condition) is often much easier than directly estimating the operator norm. The completeness of both and is essential.
The graph is a closed subspace of . With the product norm , the space is a Banach space, so is also a Banach space.
The projection defined by is a bijective bounded linear operator. By the Open Mapping Theorem, is bounded.
For any , we have , so
Therefore is bounded.
-
Differential Operators: The operator given by has closed graph, hence is bounded
-
Multiplication Operators: If , then defined by has closed graph
-
Weak Derivatives: If in and in , then (in the weak sense)
-
Automatic Continuity: Many naturally defined operators are automatically bounded by closed graph arguments
For a linear operator between Banach spaces, the following are equivalent:
- is bounded
- has a closed graph
- If in and in , then
The Closed Graph Theorem fails if either or is not complete. For instance, on incomplete spaces, there exist unbounded operators with closed graphs. Completeness is absolutely essential.
This theorem is indispensable in PDE theory, where verifying continuity of solution operators directly can be extremely difficult, but showing closed graph is often straightforward using integration by parts and weak convergence.