Study And Implement Of Automatic Test Paper Generation System Based On Genetic Algorithm

With the deepening of educational reform，the examination which plays an importantrole in the teaching activities has changed greatly, Now, more and more test papers aregenerated by the auto-generating test system. The auto-generating test system can effectivelyimprove the traditional way of examination, it also can increase the efficiency of generatingtest papers. The test paper generating is an optimized problem to multi-objective parameterwith certain restriction. Thus the efficiency and quality of this system mainly depends on thetest composing algorithm and the test database design.In order to improve the quality of the test paper, and enhance the scientific nature andobjectivity of examination, the subject of this thesis will not only change the traditional wayof proposition and build a comprehensive, scientific and rational examination questionsdatabase, but also generate test paper automatically according to teaching requirements. Usingimproved genetic algorithm to select questions, then generate a test paper.Firstly the background of the research and the composing algorithm was investigatedwidely. After a series of theoretical research, the comparison and analysis of differentcomposing algorithm, finally the strategy of generating test paper based on the GeneticAlgorithm was adopted. In this thesis, Visual C++6.0and Access2003were used asdevelopment tools, which realized an auto-generating test paper system. At last, the results offunction test and system test showed that the auto-generating test paper system based ongenetic algorithm are obviously better than the traditional algorithm both in the success rateand consume time of composing test paper. With friendly interfaces, the system greatlyimproved the efficiency of composing test.