Intelligent Test System Research And Implementation

With the Internet and computer technology developing, the field of education goes closer and closer to the computer technology, and more and more of the teaching contents become computerizable, the examination link which has played an important role in teaching and learning activities has also undergone a great change, the use of computerized examination system could significantly reduce the examination cycle, reduce duplicatie work of teachers, and improving work efficiency. Using test paper auto-generation system can enhance the quality of the papers and increase the examination of the scientific and objective.In the test paper auto-generation system, a very important issue is how the questions would be selected in accordance with the requirements of teachers and teaching work to generate desired papers. At present there has been a variety of algorithms for test paper auto-generation, such as priority strategy, random sampling strategy, retrospective testing strategies, genetic algorithms. But these algorithms often tend to fall into the local optimal point or spend too much time on the issue of huge solution space or multi-peak problems.In the text of this paper we analyse all kinds of evaluation factors, the effect of these factors and the relationship between them. On the basis of these knowledge we adopt the sum of the weighted indicators to construct fitness function of test paper auto-generation algorithm.Aiming at inefficient search and vulnerable emerging immature convergence of the basic genetic algorithm, this paper brings forward a new progressive genetic algorithm. The algorithm includes the test paper strategy, coding schemes, the determination of fitness function, select cross-mutation operator, and the realization of genetic algorithms, and uses chaos theory in cross operator and aberrance operator of genetic algorithms. Through experiments this paper verifies the global search performance, efficiency and effectiveness of the algorithm. The experiments show that the progressive chaos genetic algorithm has significantly improved in the global search performance and convergence rate.Using C# and Access database technology, this paper designs and realizes the corresponding test paper auto-generation software on the base of gradual genetic algorithm. The experimental results show that the gradual genetic algorithm is rapid and the paper quality is good, and the success rate is also higher than that of traditional genetic algorithm, so the algorithm is reasonable and effective.