Invitation to data reduction and problem kernelization

Abstract
To solve NP-hard problems, polynomial-time preprocessing is a natural and promising approach. Preprocessing is based on data reduction techniques that take a problem's input instance and try to perform a reduction to a smaller, equivalent problem kernel. Problem kernelization is a methodology that is rooted in parameterized computational complexity. In this brief survey, we present data reduction and problem kernelization as a promising research field for algorithm and complexity theory.