Minimax theorems

Many boundary value problems are equivalent to Au=O (1) where A: X ---+ Y is a mapping between two Banach spaces. When the problem is variational, there exists a differentiable functional 0 and e E X such that lIell > rand inf