Automatic generating test paper strategy based on improved Particle Swarm Optimization

The basic objective of generating test paper is to generate a high quality paper according to consumer´s requirements. The existing algorithms are likely failed because of partially fulfilling constraint conditions of generating test paper. An improved particle swarm optimization is proposed for the problem. Natural number coding and subsection strategy are adopted in the new algorithm. The results of experiment show that the improved particle swarm optimization is helpful to improve efficiency of generating test paper.