| With the rapid development of computer technology, computer assistant teaching system is extensively applied to the teaching of all kinds of schools more and more. In the computer assistant teaching system, the test system is a focus research problem, and the strategy of intelligent auto-generating test papers is the core of the test system. Its performance directly influence on the goodness and badness of the test system. At present, the strategy of intelligent auto-generating test papers is usually based on the genetic algorithm. However, traditional genetic algorithm requires the diversity of initial population and usually emerges on the phenomenon of immature convergence.Above all, this paper systemically summarizes the research background of auto-generating test paper strategy, detailedly analyzes all kinds of constraint conditions during the process of auto-generating test paper. In term of the result that the scores of examinees take on the characteristic of normal distribution, this paper presents a new method of scientifically allocating the test questions whose the ratio of both the difficulty and number of the item are different, and creates a mathematic model to solve this problem.Secondly, Partheno-Genetic Algorithm is advanced to process sequent combination optimization problem and it mainly adopt serially coding style. Aiming at the characteristic of test paper system. This method makes use of binary-coding Partheno-Genetic Algorithm and segment-gene coding style. This coding style can resolve two constraint conditions of question sorts and question quantities during coding and simplify the problem of drawing the results. At the same time, two shortcomings of traditional genetic algorithm are solved by the advantage of Partheno-Genetic Algorithm. The simulative testing result proves that Partheno-Genetic Algorithm is much better than traditional genetic algorithm.Finally, the engineering implement problems of the auto-generating test system...
|
PDF Full Text Request