Research Of Test Paper Based On GA And Realization Of Item Bank System

With the rapid development and wide application of computer and communication technology, as well as ceaseless reform of the methods of education, distance education which takes the elasticity period of schooling and the interactive teaching as the chief features, has become the main way to construct life-long education system for a modern society. Because of the separation among teachers and students of modern distance education, individual teaching evaluation according to students'personality has become an essential part to constantly improve quality of distance education. To build item bank system is the key to this problem.As the rise of standardized tests, research and development of item bank system have gradually infiltrated into the field of education. It has two advantages: on the one hand to get individual teaching evaluation according to the characteristics of students'personality, improve its interests in learning and enhance study effect; On the other hand to reuse the limited educational resources, effectively lighten the enormous workload even repeated work of teachers in process of papers and examines, and cause teachers more energy to plunge into the teaching and management, protect students'initiative, and reduce educational expenditure. And whether an item bank system is able to fully play its great role in educational and vocational personnel selection fields depends on the outstanding test paper algorithm. Therefore, this paper's proposal has its value in academic meaning and application.This dissertation deep research the strategy and related algorithms about test paper which mainly includes the following several parts:①To research three kind of automatic test paper algorithms (Based on random number test paper algorithm, Backtracking test paper algorithm and Based on GA test paper algorithm) which are commonly used in present, hold the characteristics and understand the principle and the operation process; To mainly analyze the design and process of most commonly used test paper algorithm based on GA, point its deficiency.②To research principle and operation process of genetic algorithm; To selectively analyze the influence of fitness function on GA and master features of coding and decoding of population, as well as in every processing mode of heredity operation (choice, heredity and variation).③On the basis of the analysis of the features of traditional test paper algorithms and the parameters of strategy, an advanced algorithm based on GA has been proposed. It takes knowledge points as basic test points, the count of questions as the most important goal. In stead of common linear fitness function, it considers the influence of question difficulty, difference and distribution of knowledge points to quality of papers and sets respective fitness calibration and influence factor It uses zoning codes which separates population coding from individual question coding; In order to search best solution as quickly as possible, the improved modes have been advocated based on the research of selection, crossover, and mutation.On the basis of theoretical research, the author has completed item bank system, which unifies feature of distance education. And it has been put into actual operation in DEC of Chongqing University. It gets good application and the practice has proved the algorithm can produce paper fast with relatively good quality.