WEB-based Multi-objective Intelligent Test Algorithm

With the rapid development of computer science, text without paper and online text come to us. The mess task of making up test and huge amount of homework in online study are a lot. The intelligent test-paper composing system has a significant meaning in solving this problem.This thesis mainly investigates relative knowledge of intelligent test-paper composing system and the present condition and development of intelligent test-paper composing skills and compares several popular algorithms in and abroad. Research the key algorithm and skill of pretreatment to test database, relate to the on line skill and database design. Introduce the multi-target skill and analyze the evaluation system of intelligent test-paper composing system. In the mathematic model introduce the constraint condition and error analysis and the function and meaning of objective function. Research the theory and mathematic model of improved particle swarm optimization algorithm which is chosen to be the key algorithm because of its randomness in choosing question in the early period, with a solution in premature convergence mechanism and suits small and mid-size intelligent test-paper composing system in time and space the algorithm works. Compare the preliminary classification algorithm and nearest neighbor classification algorithm. In the last, introduce the design of overall framework, realizing of the key model, design of the database and the analysis of the date of the intelligent test-paper composing system.In all the research, we obey the principle of satisfying the practical requirement of teaching, design and realize the algorithm and database. The system is realized by the database of SQL Server2005, and Java language.