Study And Development Of Online Examination System Based On J2EE

It rapidly promotes development of computer remote education that computer networks and multimedia technology are widely popularized in a modern society in various fields. However, with constantly increasing of teaching and learning resourse and a growing number of people taking part in remote education, remote education system is demanded more powerful function. As an important part of remote education system, an online examination system has an important influence on a fit and unfit quality of remote education system. Therefore, studying and developing of a powerful, and scalable and easy to maintain online examination system has a very important practical significance.An online examination system is designed and achieved based on Struts framework in order to offer students online test in the paper.The sharement and the quick dispense of praxis resources are achieved in the examination system, and online test is offered to students taking part in remote education.Struts three-tier based on Model-view-controller is used in the examination system. View is responsible for interaction with examinee; Model is responsible for processing business flow, business state and making business rules; Controller is responsible for receiving request from View and Model and relating between View and Model. The implementation of three-tier structure enables the system to have distinguished quality of flexibility, currency, and compatibility.Genetic Algorithm is used in the auto-generating test paper system in order to achieve the auto-generation test paper system intelligentizing, and the traditional Genetic Algorithm is improved at some aspects. NaturalNum-code replaced Binary-Code in order to improve the traditional Genetic Algorithm at chromosome-Code in the Genetic Algorithm. The questionID is directly used as Cheomosome-Code in order to save the time of coding and Encoding. It maintains multi-seeds effectively, and avoids to run into trap of the part best value and precocious convergence that Simulated Niche is introduced in the selection operators. It improves the crossover probability and mutation probability, enables the Genetic Algorithm to find the optimizing crossover probability and mutation probability to suit itself, improves convergence speed of the auto-generating test paper system, and avoids running into precocious convergence that self-proper theory is used in the Genetic Algorithm.