Abstract:
Problems that reduce to consistency or inconsistency of systems of linear equations or inequalities arise in many divisions of theoretical informatics. The examples are problems in linear programming, machine learning, multicriteria optimization, etc. There exist different stability measures for the property of consistency or inconsistency, and different information constituents are possible (all the input parameters, the coefficient matrix, the vector of constraints). In this paper, variations of all parameters are examined in combination with an additional constraint important in applications, namely, the nonnegativity of feasible points.
Key words:matrix correction, inconsistent systems of linear equations and inequalities, stability of systems of linear equations and inequalities.
Fulltext:
PDF file (233 kB)