Privacy Protected Collaborative Optimization Algorithm For The Cross-organization Constructed LP Model

In the context of economic globalization and the rapid development of information technology, more and more companies recognize that cooperation between enterprises may be more important than competion in a highly competitive market economy. For example, In supply chain management, companies or organizations generally focus on the total production or profit of the entire supply chain when considering to improve their competitiveness and customers’satisfaction, namely supply chain collaborative decision making. Supply chain collaborative decision making is based on the enterprises or organizations’ information sharing. Although the supply chain members have the same economic interests, there are also conflicts of interest between them. Information sharing may(or have to) cause leakage of confidential information of enterprises and bring the negative influence to them. Therefore, enterprises seldom share information with other members in actual production activities.So it is difficult to reach the global optimal objectives of the entire supply chain. There is important significance in studying how to reach global objectives in collaborative decision making without disclosing the participates’ privacy information. And how to achieve the global optimization in case of no privacy infprmation leakage is also a hard problem in supply chain management.Secure multi-party compution(SMC) is an important tool to solve this kind of problems. Secure multi-party computation mainly solves the collaborative computing problems between parties who do not trust each other. SMC can guarantee parties’independence of their input and correctness of their computing results with no privacy information leakage. In this paper, on the basis of secure multi-party computation theory as well as the basic protocol of secure multi-party computation, following work have been done:1) The development of work on distributed collaborative optimization with no information privacy disclosing, SMC and application of SMC in domestic and abroad have been discussed in this paper.2) An efficient approach based on random matrix transformation to the linear programming model of which constraint matrix data is distributed in different parties by rows and objective function is also another party’s privacy information proposed in this paper.3) Also this paper proposed an efficient protocol for arbitrary distributed K-LP problems based on D-W decomposition and column generation algorithm; and the communication and computing costs of the protocol have been analyzed.4) Finally, we creatively combined SCG protocol and game theory and proposed an incentive compatible protocol for K-LP under malicious model.