Research On Intelligent Test Paper Composition System Based On Genetic Arithmetic

A paper is reasonable, if people want to reach a volume value is a measure of whether a set of intelligent test system of standards, along with computer technology in the wide application of artificial intelligence, many experts and scholars turn their attention to the smart set volume of the system, in this system the most attractive for the researchers is to model and test paper test paper algorithm. This is the test paper through the traditional method exists "test paper slow, low success rate, test paper quality is not high"and other shortcomings, proposed genetic algorithm is applied to test system of the advantages.The article analyzes the evaluation of examination papers which indicators, the indicators have little effect, the existence of the indexes what kind of relationship, according to the analysis of the papers build a distribution model, and then analyzes the process of generating questions papers should be what kind of constraints. Select the appropriate strategy for comparing the test paper and create test paper model, then the model is simplified. The mathematical model of the test paper, the second describes the use of dynamic hierarchical genetic algorithm to solve the problem of the step test paper, which covers the key technologies which are:the establishment of simulated test bank, select the test paper strategy, determine the coding scheme to determine fitness function, select the crossover and mutation operator, dynamic hierarchical genetic algorithm realization.Finally, the general genetic algorithm and improved genetic algorithm for generation of paper compared the experimental parameters to determine the improved genetic algorithm is feasible and reasonable.