Implementation of the Gale-Shapley algorithm for automating reception of university applicants

Introduction. The article considers the realization of the algorithm enrollment of applicants to university on the basis of the Gale-Shapley algorithm with various parties initiating. The aim of this study is to analyze the performance and stability of the computer implementation of the algorithm. The qualitative and quantitative differences received on the same data sets are discussed. Results. The authors propose two options algorithm for admission entrant, which differ in side initiation of the "enrollee" or "university". For the algorithm, with a side of initiation "enrollee", allocation can be made on specialties in process entering data about entrants, thereby to maintain the latest information about placing, but it's worth noting that if you delete or change data of entrants you must run the algorithm from the beginning using the already entered data about the applicants. And for the algorithm with the initiating party "university" is possible to prioritize enrollee in specialty in process filing their applications and run the algorithm if necessary. It is worth noting that the resulting placement will be more beneficial to initiating parties. Also discusses the characteristics of the average execution time when these values change: the number of applicants, the quota in the specialty, the number of priorities in the entrant, the number of specialties. The authors point to the differences results of the algorithms with different parties initiating that is not only the qualitative difference of the results, but also quantitatively. The quantitative difference is due to "equality" of entrants a relatively some specialty. Discussion and Conclusions. Variants proposed by the authors can be used for automating enrollment of university applicants.