Selection of Control Tasks for Students Using Neural Networks and Multi-Criteria Optimization

One of the most important parameters of a student’s progress is the correct selection of control tasks when conducting an intermediate cut of knowledge. This paper proposes the optimal compilation of test tasks, taking into account their complexity, duration of execution, the number of questions in the task, the number of topics covered and the actions required to complete the task. Since each student is individual, the question arises of the correctness of providing the same control task to students with different mental and psycho-emotional characteristics. Therefore, the optimal compilation of tasks for a particular student is quite relevant. This study will improve the student’s progress and success in general, and will also remove a number of responsibilities from the teacher related to the generation of topics and tasks when compiling the test. The paper proposes to develop a hybrid system of multi-criteria optimization with two fully connected neural networks, which allows determining the most suitable model for compiling a task of control work based on a number of features of a particular student.