Intelligent Generating Test Paper Scheme for English Class Based on Genetic Algorithm

This paper mainly discussed the implementation of genetic algorithm on generating test paper for English class. The existing genetic algorithm is optimized through the analysis on principles and key technologies of standard genetic algorithms. We use matrix encoding to improve the coding method. It can storage various properties of indicators and fitness of individual papers and improve computing speed. By the improvement of genetic operators, we propose a sectional multi-point strategy for mutation operator. A large proportion of hybridization and mutation are adopted to maintain the diversity of population, which avoid the knowledge conflict in generating paper and rapid shrinking when searching space. It also raises the speed of approximating to the global optimum and improves the comprehensive performance. Simultaneously, to ensure the best individual parent enter the offspring, we adopt the selection method based on fitness sorting during copy, to make the selection on the generation be the best strategy. The experiments results verify the effectiveness and practicality of our algorithm.