No label defined (Q3972866)

This paper is concerned with inequalities which involve abstract operators, briefly: operator inequalities. We will mainly consider operator inequalities which originate from operator equations \[ Mu=r, \tag{1} \] where \(M:R\to S\) is a given (linear or nonlinear) operator, \(r\in S\), and \(u\in R\) denotes the solution to be determined. The relation \(Mu\leq r\) (in an ordered space \(S\)) is a very simple example of such an inequality. One of our main objects is to obtain information on a solution of an operator equation (1) by solving operator inequalities related to this equation. The theory, however, allows also other applications, for instance, to eigenvalue theory. Sections 1, 2, and 3 deal with (a priori) inclusion statements. The various statements considered can all be written in the form \[ Mu\in C \;\Longrightarrow\;u\in K, \tag{2} \] where ``\(u\in K\)'' represents the inclusion statement desired, and \(C\) is a set described by inequalities involving \(M\), i.e., operator inequalities. A statement like (2) will also be called a range-domain implication (for \(M\)). A simple example is the statement \[ M\varphi\leq Mu\leq M\psi\;\Longrightarrow\;\varphi\leq u\leq\psi, \] which holds for inverse-monotone operators \(M\) in ordered spaces. By applying (2) to the operator equation (1), one obtains the following statement: \[ r\in C\;\Longrightarrow\;u\in K \text{ for each solution of } Mu=r. \tag{3} \] Section 4 treats existence and inclusion statements of a corresponding form: \[ r\in C\;\Longrightarrow\;\text{there is a solution }u\in K \text{ of } Mu=r. \tag{4} \] If a statement of the latter form holds, one can, in theory, carry out an existence proof by verifying the inequalities which the premise ``\(r\in C\)'' represents. In section 5 we will show that, for certain problems, the verification of all assumptions required for an existence proof can be reduced to the construction of two algorithms: an algorithms for calculating approximate solutions, and the algorithm for calculating (constant) bounds for certain elements. In addition, we will briefly report on a corresponding program for two-point boundary problems. For the entire collection see [Zbl 0771.35035].