协同优化决策中数据水平分区的隐私保护算法研究

在跨组织协同优化决策问题中的参数来源于不同主体的数据.在缺乏可信第三方时难以完成全局优化问题的求解.本文运用随机矩阵转换和加密技术方法来解决约束条件中数据水平分区优化问题的协同计算，克服了扰动或差分算法对问题结构以及解结构潜在的不稳定影响.提出的安全协议一方面可以保证在保护各方隐私信息的前提下得到计算结果与集中式的结果具有一致性，另一方面也具备良好的防推理攻击能力.该研究可广泛应用于供应链系统或企业联盟间的决策优化问题的协同安全计算问题.

Privacy-Preserving Algorithm Research on Horizontally Partitioned Data of Collaborative Optimization Decisions

In the cross-organizational collaborative optimal decision-making, the parameters of optimization usually originate from different subject data. It is difficult to solve the global optimization problems due to the lack of a trusted third party. A method combining the random matrix transformation and encryption technology is put forward to solve the collaborative optimization of the horizontally partitioned data. This method also overcomes the potential destabilizing impact of structural problems and structural solutions when using disturbance or difference algorithm. On one hand, the security protocol can ensure the consistency of calculation results with privacy-preserving and centralized results. On the other side, it can prevent the potential inference attack. The study can be widely used in collaborative computing security issues of optimization decision problems among enterprise alliance or supply chain.