Definition:

filename:   name of the instance file

dim: structure of the table

n1:  number of categories in first dimension, not including "total"

n2:  number of categories in second dimension, not including "total"

n3: number of categories in third dimension (if any) not including "total"

n4: number of values in fouth dimension (if any) not including "total"

cells: total number of cells (including internal, marginals and totals)

sums: total number of equations linking the cells

cells/sums: ratio of "cells" divided by "sums"

|P|: number of Primary suppressions (i.e., the number of given sensitive cells)

|Pl|: number of given non-zero protection levels (before the preprocessing of the algorithm), typically 2 for each sensitive cell

l1u1: number of sensitive cells with both upper and lower protection level after preprocessing

l1u0: number of sensitive cells with only lower protection level after preprocessing

l0u1: number of sensitive cells with only upper protection level after preprocessing

l0u0: number of sensitive cells with none protection levels after preprocessing

pl0: number of non-zero protection levels after preprocessing, typically smaller than 2 for each sensitive cell.

tprep: time (in seconds of a Pentium III/866 Mhz) of the preprocessing

theur_: time (in seconds of a Pentium III/866 Mhz) of the initial heuristic procedure

UB_: total loss of information of the initial feasible solution proposed by the initial heuristic

theur: time (in seconds of a Pentium III/866 Mhz) of the initial heuristic

UB0: total loss of information of the feasible solution found at the end of the root-node in the branch-and-cut search

LB0: lower bound of the minimum lost of information at the end of the root-node in the branch-and-cut search

time: time (in seconds of a Pentium III/866 Mhz) of the preprocessing

opt:   minimum loss of information (optimal value); see table for an optimal suppression pattern

SUP |S|:  number of Secondary suppressions in an optimal suppression pattern

time: time (in seconds of a Pentium III/866 Mhz) for finding an optimal suppression pattern

BestLB: best lower bound of the minimum lost of information if the search quits before the end

nodes: number of explored branch-and-bound nodes

Ttime: Total computing time for the whole branch-and-bound search

Tsepar: time for finding violated inequalties

Toptim: time consume by the LP-solver (e.g., CPLEX 7.0)

cuts: total number of violated inequalities introduced during the "separation procedure"

pool: maximum number of inequalities saved in the pool-data structure

bender: number of Benders' cuts

capa: number of capacity cuts

bridge: number of bridgeless cuts

cover: number of conver cuts

cover2: number of 2-cover cuts

gomo: number of Gomory cuts

comb: number of comb cuts

path: number of path cuts

path2: number of 2-path cuts

iterations: total number of calls to the "separation procedure"

UBtype: "H" when the optimal patter is found by the heuristic approach; "L" when it if found by the cutting phase

Taudit: time for solving all the attackers problems to find the interval ranges for all suppressions in the proposed pattern

Ttime+Taudit: total time for protecting a table and checking the auditing phase.

|PuS|: total number of (primary and secondy) suppressions (i.e., missing values) in the proposed pattern

Add WIDES: sum of the interval ranges under the suppressions in the proposed pattern