Privacy preserving Cloud-based Quadratic Optimization

This work proposes a protocol for privately solving constrained quadratic optimization problems with sensitive data. The problem encompasses the private data of multiple agents and is outsourced to an untrusted server. We describe the desired security goals and investigate the information leakage from duality theory. We present an interactive protocol that achieves the solution of a strictly quadratic convex optimization problem with private linear cost and private linear inequality constraints, by making use of partially homomorphic cryptosystems to securely effectuate computations. Then, we provide extensions to the protocol in order to also consider equality constraints and to obtain a speedup of the performance.