Abstract
Protecting tabular data often requires secondary cell suppression: once sensitive cells are hidden, further cells must be suppressed so the sensitive values cannot be recovered by arithmetic — an NP-hard optimisation problem. This paper benchmarks commercial and free/open-source mathematical-programming solvers on the secondary cell suppression problem, comparing them on solution quality and computational performance. The results give practitioners evidence on whether open-source solvers can match commercial ones for protecting official tables, an important question for agencies with limited budgets. Published in Transactions on Data Privacy, the findings informed the tabular-protection tooling developed in the sdcTable ecosystem.
statistical disclosure limitation secondary cell suppression linear programming