Parallel strategies for heuristic optimization algorithms: Extracting multi-choice tests

Abstract

The main goal of this thesis is to study parallel strategies and heuristic optimization algorithms to solve optimal extracting multi-choice tests from question banks with objective difficulty. We analyze and evaluate efficiency methods for problems of extracting tests by experimental results analysis. For such a task, we cover three aspects of extracting multi-choice tests, which are described as follows: First, we study heuristic optimization algorithms to extract multi-choice test to satisfy levels of difficulty requirement (single-objective optimization problem). Second, we focus on investigating parallel strategies for heuristic optimization algorithms to extract multi-choice k tests to satisfy predefined levels of difficulty requirement by users. (multi-swarm optimization problem). Finally, we consider parallel strategies for combination heuristic optimization algorithms to extract multi-choice k tests to satisfy predefined both levels of difficulty requirement and required time test by users (multi-objective optimization problem). The main result of the thesis generated proper methods for extracting multi-choice k tests based on multiple redefined requirements by the users, and the tests have the same difficulty level to ensure fairness for all students when they join the same exam. The studies support well the evaluation of students’ study progress in the education sector. Which are suitable for application in both regular tests and final exams, etc.