Research And Implementation Of Online Examination System Based On Genetic Algorithm

In the context of the rapid development of the Internet, the traditional teaching model cannot meet the needs of modern education any more. Thus the online teaching has become the inevitable trend of education reform. The degree of the development of the online examination system is closely related to the real application prospect of network teaching. Therefore, in order to improve network teaching, it is critical to take intense research into promoting the normalized and scientific online examination. The intelligent test paper composition is a parameterized calculation issue which is constrained by multiple targets. Its performance is largely depended on the design of item bank and the efficiency of the algorithm. The current online exam system has mainly realized the function of test-paper automatic-organizing, online exam and automatic marking. But the efficiency of the algorithm is not very high, the quality of test paper composition can’t meet reality demand and the function of the system is not perfect.On considering of the drawbacks of the current online test system, this thesis took further research into the intelligent test paper composition and accomplished the math modeling towards it on the basis of Secondary Constructor Continuing Education Platform. Then this thesis proposed a genetic algorithm based on the standard deviation to optimize the genetic-operation of test paper composition. Finally, to realize the secondary registered construction engineer examination subsystem and operation, the improved algorithm is verified the experimental results which show that the research has certain reference value for practical applications. In this dissertation, the main work and innovations are as follows:(1) In order to solve the problems of the low quality of papers, we regulated the each attribute constrain of the test and implement digitized management to the question bank. We combined it with genetic algorithm and designed the fitness function. What’s more. I mapped the various attributes to the expressing space of the genetic algorithm and constructed the mathematical modeling.(2) In view of the low efficiency of the genetic algorithm, we analyzed the existing adaptive genetic algorithm and proposed an optimized scheme using standard deviation to improved mutation operator and crossover operator. Finally the adaptive genetic algorithm was improved. I also used two functions to test the improved algorithm. The simulation results show that the improved algorithm perform better than traditional adaptive algorithm.(3) On considering of the bottlenecks of genetic performance, we made improvements towards coding scheme and the genetic operator.I designed a real-number-coding scheme based on topic section and reformed the traditional roulette selection operator.I proposed the backbone retention strategies to improve the performance of the system.Finally, the thesis tested the effects of the optimized algorithm and its optimized operating process on the basis of the implementation of Secondary Constructor Continuing Education Platform of ShanDong Province. I also tested and analyzed the whole system.