Solving a system of linear inequalities for generating test data

摘要

Data Flow Analysis can be used to find some of the errors in a computer program and gives as output a set of dubious paths of the investigated program, which have to be checked for executability. This can be done by solving a system of inequalities in order to obtain input data for that path. This article discusses how to obtain a reliable solution of this system in the linear case, when rounding effects are taken into account. The method is based on the simplex algorithm from linear programming, and returns a solution in the middle of the feasible region. The general nonlinear case is much more difficult to handle.