Castro J., González J.A. (2010) A Tool for Analyzing and Fixing Infeasible RCTA Instances. In: Domingo-Ferrer J., Magkos E. (eds) Privacy in Statistical Databases. PSD 2010. Lecture Notes in Computer Science, vol 6344. Springer, Berlin, Heidelberg
Minimum-distance controlled tabular adjustment methods (CTA), and its restricted variants (RCTA), is a recent perturbative approach for tabular data protection. Given a table to be protected, the purpose of RCTA is to find the closest table that guarantees protection levels for the sensitive cells. This is achieved by adding slight adjustments to the remaining cells, possibly excluding a subset of them (usually, the total cells) which preserve their original values. If either protection levels are large, or the bounds for cell deviations are tight, or too many cell values have to be preserved, the resulting mixed integer linear problem may be reported as infeasible. This work describes a tool developed for analyzing infeasible instances. The tool is based on a general elastic programming approach, which considers an artificial problem obtained by relaxing constraints and bounds through the addition of extra elastic variables. The tool allows selecting the subset of constraints and bounds to be relaxed, such that an elastic filter method can be applied for isolating a subset of infeasible table relations, protection levels, and cell bounds. Some computational experiments are reported using real-world instances.
statistical disclosure control controlled tabular adjustment mixed integer linear programming infeasibility in optimization elastic constraints elastic filter