Research On Privacy-Preserving Protocols For Selection

Secure Multi-Party Computation (SMC) has been a hot research topic in the field of information security in recent years. In a distributed data environment, participants are expected to coordinate their efforts in completing a computational task. If every participant needs to provide his input data, and he is not willing to let other participants know his input, we call such an input private input. The multi-party protocols mandated by such an input is Secure Multi-Party Computation (SMC). The difference between secure multi-party protocols and traditional cryptography protocol is that SMC does not consider the external attack, and it works under a secure framework for information, and focuses on solving the problem of the "leaks" within the parties.Privacy Preserving Statistical Analysis (PPSA) is an important topic in SMC. If the basic data to be processed in a statistical analysis come from two or more participants, and the participants are unwilling to reveal their data directly to other participants or a third party, then it is time for the Secure Multi-Party Computation to step in. At present some preliminary conclusions have already been reached in this field.The secure multi party computation of median and the kth element is a special case of PPSA. If we want to compute the common kth element of the data possessed by two or more parties, and prevent the private data of each party from being known by the other parties, we must design a secure multi-party computation protocol.This thesis is based on the studies and researches of the secure multi party computation of median and the kth element. After we had studied the literature on the computation of median and the kth element, we explored the different situations under the need for privacy, and made some achievements under the support of the basic protocol of secure multi party computation.1. We come up with a very simple but effective conclusion on rapid computation of multi-party median. We find that for multi-sequence of numbers, if we delete the same number of elements before and after the median, the median of the new sequence remain unchanged.2. We proposed a protocol for securely computation of the median. Based on the logic and steps used in the computation of multi-party median, and the secret comparison protocol, through several rounds of interaction, we have solved the problem of secure computation of the median for two party and three party against the prerequisite for private input..3. The median is a special case of the kth element. If we can get the kth element of the parties concerned at will, we shall be able to get the maximum, the minimum, the median and the complete sequence. We modified a protocol proposed by Gagan, we optimized the marginal conditions and specified the medians.4. We proposed a protocol for securely computation of the kth element of the multi party sequences, by using the "Counting" method and ring style secure multi party sum protocol which was proposed by us.