Research On The System Of Generating Test Paper Based On Subsection Integer Encoding Adaptive Genetic Algorithm

Intelligent test paper generat ion system is a key research topic in the rapid development of university informat ion construction in recent years.Intelligent test paper system can reduce the teacher repeat ed workload,improve the quality of papers,and increase the objectivity of the examination.The algorithm of Intelligent test paper generat ion is a multi-object ive optimization problem and it is the core content of intelligent test paper.Because of its good convergence and high robustness,genetic algorithm is very advantageous to so lve this kind of problem.The tradit ional genetic algorithm is "premature" and easy tofall int o the local opt imal so lut ion and so on,which affect the convergence of the algorithm.Is proposed in this paper by subsection integer coding and adapt ive operator method of geneticalgorithm is improved,and based on the improved algorithm to design an intelligent applicat ion system.The main research contents are as follows:(1)The improvement of genetic algorithm.Using piecewise integer coding method of generat ing system of chromosome will be of the same type to a segment,each segment of code do not affect each other,each code segment independent ly genetic calculat ion,the algo rithm set up the target attribute weights to construct the adaptat ion degree function,in the test run in the process of using the adaptive crossover operator and mutat ion operator,selected to adapt to the environment of the highest individual and the individual does not participate in cross genetic operations such as co mput ing,the worst individual will be replaced directly,enable outstanding individual can continue to ensure test paper quality and success rate,extent improves the running efficiency of the algorithm.(2)The design of intelligent test paper system.it implements a system of generat ing test paper based on subsection integer encoding adaptive genetic algorithm,the system uses B/S structure,the system is divided intofront and back office management funct ion and the front includes online testing,answer browsing and score funct ion module,the back includes the login,course management,test library management,paper management,user management functional modules.The improved genetic algorithm has a free combination of individuals,which makes the informat ion search as an independent way,without relying on any knowledge.Algorithm will also be based on the environment,derived fro m the corresponding offspring,the offspring of environme ntal adaptabilit y is relat ively strong,so as to adapt to the changes in the database information.In the aspect of parallelism,the problem of large amount of computat ion is effect ively so lved.It is the main characteristic of the informat ion group tofind the optimal individual and tofind the individual.Finally,in "C language" course as an example test paper 50 times test,improved genetic algorithm the number of successful 46 times,average time of 25 seconds,compared with other three kinds of tactics of the test paper,test results show that the improved algorithm,the success rate and the convergence speed is improved obviously,has better practicability.