中文摘要：

为了降低资源受限用户求解凸二次规划问题的计算量,提出了可验证安全的凸二次规划外包计算协议。新协议首次引入置换技术,将原始问题盲化转换成随机问题,然后外包给云服务器求解,最后验证服务器返回结果,减少了用户端的计算量。安全性分析表明,在完全恶意模型下,新协议可以保证输入输出数据的隐私性,且能以最优的概率检测出云服务器的不诚实行为。仿真实验表明,与现有协议相比,新协议中用户在转换和验证阶段所需时间明显降低。

英文摘要：

To reduce the computation required for resource-constrained clients when performing convex quadratic programming,we propose an outsourcing computation protocol for convex quadratic programming whose security can be verified. In the new protocol,the client first utilizes a permutation technique to transform the original problem into a new random problem,which the cloud server receives and solves,and the client then verifies the returned results. Thus,the new protocol can reduce the client＇s amount of required computation. Security analysis shows that the proposed protocol can protect the privacy of the input and output data,and detect any misbehavior by the cloud server to indicate the probability of a malicious model. Experimental results show that the new protocol has a comparative advantage over existing protocols in its transformation and verification efficiency.