Computational Complexity and Feasibility of Data Processing and Interval Computations

The input data for data processing algorithms come from measurements and are therefore not precise. It is therefore necessary to estimate the accuracy of the results of data processing. Even for the simplest data processing algorithms, this problem is, however, intractable. This book describes for what classes of problems interval computations (i.e. data processing with automatic results verification) are feasible, and when they are intractable. This knowledge is important for algorithm developers, because it will enable them to concentrate on the classes of problems for which general algorithms are possible.