Research on Intelligent Test Paper Generation Based on Improved Genetic Algorithm

In order to avoid slow-convergence and local convergence of simple genetic algorithm (SGA) for intelligent test paper generation, a kind of improved genetic algorithm (IGA)has been proposed in this paper. This algorithm uses unceasing elimination of similar individual method to quickly enlarge the search space and to stabilize the individual diversity of the group. Experiment results show that the test paper formed by the algorithm meets all the user requirements if the quantity of test questions is moderate and reasonable.