The Research Of Computerize Adaptive Test Based On Integer Sub-encoding Genetic Algorithm

With the development of information and technology, people start to doubt some traditional ways and methods for dealing with affairs. And, some of them attempt to change and improve the traditional things by means of the modern technologies. In this case, the revolutionary changes have been taken place to the testing ways for evaluation of personal ability. The traditional examination (such as paper, pen, etc.) is being replaced by the test system on the computer network. Clearly, it is necessary to develop the high level self-adaptive test system.The generating system for intelligent test paper is the core of a self-adaptive test system. The shortcomings of traditional ways for generating test paper are much time cost and poor performance. The main reason is that the traditional algorithms are based on binary encoding. In this thesis, after analyzing the traditional algorithm for auto generating test paper, an improved genetic algorithm based on integral number encoding was proposed. The performance of this new algorithm is promising.The content of this thesis is organized as follows: first, the related theories to self-adaptive test system and the traditional algorithm based on genetic algorithm were presented; then, an improved algorithm for auto generating test paper was shown; finally, the detailed design of a self-adaptive test system was contributed.