The Study And Application Of Automatic Generating Test Paper System Based On Genetic Algorithms

As the development of computer technique and Artificial Intelligence, the research of test paper auto-generating system is paid attention to by more and more scientists gradually. It not only refers to the establish of test paper auto-generating mathematics model, but also includes the application arithmetic. Automatic generating test paper is to generate a test paper with specific total point, total time, type, quantity, knowledge, difficulty and so on according to a certain requirements by selecting test questions from item bank system automatically. The test paper auto-generating is an optimized problem to multi-objective parameter with a certain restriction. The optimization is implemented very difficultly by traditional mathematics method. The quality and efficiency of automatic generating test paper is determined by the design of test questions database and select test questions algorithm absolutely. Since automatic generating test paper needs to satisfy consumer's variable demands with great randomness and rationality, a more effective algorithm is in great demand.Above all, we give a brief introduction to the history of computer test paper generation technique development. On the basis of analyzing a large quantities of references home and abroad, this thesis has made a survey on the automatic generating test paper algorithm. Then we analyzed the shortcomings of the techniques. Based on that, we bring forward major difficult problem in automatic generating test paper technique and show the arduous task ahead of us.Secondly, this thesis summarizes the research background of auto-generating test paper systemically, detailedly analyses of evaluating indicators and their functions, relationships of some important indicators; a test paper generation mode was set up with the distributions of the evaluating indicators. And we analyse all kinds of constraint conditions during the process of auto-generating test paper farther. We establish a model of generating test paper problem and simplify it. We obtain a standard process of test paper generation by analysing mathematical models in test paper generation problem. Based on that we study according to the generation arithmetic which are in common use now. We give our best efforts on a detailed introduction to the generation strategy based on genetic algorithm, both theoretically and practically.Thirdly, aiming at the characteristic of automatic generating test paper system., this thesis makes use of the national code which is used in the genetic algorithm in order to decrease space of chromosomes.This coding style can resolve two constraint conditions of question sorts and question quantities during coding and simplify the problem of drawing the results. The initialization test paper will be selected two to match randomization, using the condition" order crossover", and one-point crossover mutation of the same topic group in the chromosome crossover and mutation improved.Finally, the engineering implement problems of the automatic generating test paper system based on genetic algorithms are discussed in detail, and the system has been achieved using C#. In this thesis, some codes are provided.Since the test paper generating system refers to lots of theories itself, the author mainly use Genetic Algorithm to research a algorithm to settle the questions of automatic generating test paper system. Aiming at the questions of automatic generating test paper system , in order to verify the validity of automatic generating test paper based on GA, we made a lot of experiments on auto generating test paper, the result of experiments shows that the method is reasonable and effective. Because of the importance of algorithm on intelligence teaching system theory research and fact applying, our research is necessary, it is efficient to overcome the existing problems and has considerable theoretical and practical meaning.