Design And Implementation Of Intelligence Testing-Paper Generating System Based On General Item Bank

The developments of the computer network technology, the computation technology as well as the artificial intelligence technology impel the progress of computer test system continuously. The computer test system has abandoned the limitation of the traditional test pattern and has brought the convenience for many examinees, simultaneously reduced teacher's redundant work. The intelligence testing-paper generating system (ITPGS), which is the key part of computer test system, has become the hot research area in the fields of pedagogy and computer science.The essence of ITPGS is solving the optimization problems of multi-objective parameters which will impact the quality of examination paper. In this thesis, the improved objective function and the mathematical model of ITPGS have been proposed firstly. By introducing the weight for each evaluation index, the new model converts the multi-objective optimization into a single-objective optimization. In order to determine the wights, the multi-objective evaluation decision analysis method is applied.Secondly, an improved self-adaptive genetic algorithm has been peoposed to solve the mathematical model of ITPGS. Compared with the existing algrithms, the new strategy can obtain the more quickly convergence speed and find the globally optimal solution effectively and efficiently.Then, the Calculus examination system-HEBITPGS-is designed and implemented by using the JSP programming language and SQL database technology. ITPGS is the key part of HEBITPGS. Finally, the experimental simulations have been carried out in order to validate the feasibility and effectiveness of HEBITPGS. The experimental results show that the new model can improve the quality of examination paper and overcome the premature convergence in some extent.