Privacy-preserving Algorithm for Linear Programming with Vertically Distributed Data
Guoping He · Journal of Shandong University of Science and Technology · 2011
In a privacy-preserving algorithm for linear programming with vertically distributed data,Mangasarian employed a random matrix with which the original linear programming could be transferred into a secure linear programming.However,when the random matrix is irreversible,the original linear programming and the secure linear programming are not equivalent.Here,a reversible random matrix is used to make sure that the original linear programming could be transformed into an equivalent secure linear programming.The numerical experiments show that the results obtained with this algorithm are close to the original linear programming,and with the increasing of parameter λ,the accuracy of this algorithm is also be improved.