Abstract
This paper proposes the use of multi-swarm method in particle swarm optimization (PSO) algorithm to generate multiple-choice tests based on assumed objective levels of difficulty. The method extracts an abundance of tests at the same time with the same levels of difficulty and approximates the difficulty-level requirement given by the users. The experimental results show that the proposed method can generate many tests from question banks satisfying predefined levels of difficulty. Additionally, the proposed method is also shown to be effective in terms of many criteria when compared with other methods such as manually extracted tests, random methods and PSO-based methods in terms of execution time, standard deviation, the number of particles per swarm and the number of swarms.
Highlights
Speeding up the use of science and technology to education is inevitably demanding for sustaining the development of society and addressing its cultural and educational needs
This paper proposes a parallel multiobjective particle swarm optimization (MOPSO) based on consumer-level Graphics Processing Units (GPUs), which, to our knowledge, is therst approach to optimizing multi-objective problems via particle swarm optimization (PSO) on the platform of GPU
We present an approach to extract an abundance of tests with equivalent levels of di±culty and approximate the specic di±culty-level requirement given by the user based on question banks and the parallelism of multi-swarm method in PSO algorithm
Summary
Speeding up the use of science and technology to education is inevitably demanding for sustaining the development of society and addressing its cultural and educational needs. Tests need to evaluate the correct characteristics of educational quality, so one has to standardize the multi-choice question banks to ensure the accuracy of the level of di±culty of questions. Culty depending on the composition of bank questions This makes extracting a capability test with the required di±culty using the traditional methods di±cult and it cannot be performed with the huge search space of the question bank. This problem is important when extracting multiple tests with similar di±culty for multiple students at the same time For these reasons outlined above, we keep on studying to come up with solutions for these cases. This paper solves the problem of extracting multiple-choice test based on assumed objective levels of di±culty.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
