Development of a searching algorithm based on neural networks for the optimal University studies schedule.

By the author of the work, part of the requirements for the schedule of training sessions in higher educational institutions are formalized and written in the form of equations and inequalities. All these relations are linear or quadratic expressions with integer (more precisely, Boolean) arguments. In [1], a possible formalization of the problem of finding the optimal schedule is presented, which also turned out to be an integer quadratic Boolean programming problem without additional constraints. There are methods for solving this problem [2], but they have a relatively low rate of convergence, which, taking into account the dimension of the real problem of forming the SS, does not allow using these methods to solve the problem of finding the optimal and acceptable SS. Therefore, it is necessary to develop a method for solving the obtained formalized problem, which would make it possible to obtain a solution to the problem under consideration in an acceptable time, taking into account the capabilities of modern computer technology. In this paper, we propose an algorithm for solving the problem of composing SS based on the adaptation of the classical gradient search method. In relation to the task of compiling SS, there is a large list of works, including classical monographs [3], [4]. [ABSTRACT FROM AUTHOR]