Algorithm for Optimization of the Content of the Training Course Practical Part Using the Artificial Immune System

Abstract

The article presents a model of a training system using the artificial intelligence methods for optimization of certain educational process components. The training system allows the teacher to create and optimize training courses based on the accumulated statistical information. For development of the training system, a hybrid system was chosen combining the advantages of various technologies that allow solving each problem in the optimal way. An algorithm for optimization of the content of the training course practical part using the artificial immune system has been considered. A set of the class’s practical tasks is divided into classes of tasks of similar complexity aimed at achieving similar objectives of the course. The objective function and problem limitations are formulated using H. Markowitz’s model. One of the problem’s objective functions minimizes the correlation between the complexity of tasks of different classes, which allows excluding presence of many single-type tasks in the collection of practical tasks; another objective function maximizes the effectiveness (notion “effectiveness” is introduced in the article) of the collection of tasks. The model’s variables are shares of the total number of tasks selected from each class. For optimization of the given model, a set of Pareto-optimal solutions of a bicriterial problem is found, which allows selecting the optimal relation between the tasks diversity and their effectiveness. The work offers an algorithm for finding the solution of this problem, modified for the artificial immune system. The algorithm suggested in the problem allows obtaining, in a relatively short time, a satisfactory approximation of the Pareto-optimal set for solution of the problem.