Graduate Student Re-examination Evaluation System Design Based On Fuzzy Theory

With the constantly expanded enrollment of graduate students, how to select the suitable candidates from lots of examinees becomes a big challenge for every university and college. Because the treatment of interview result which is mainly based on manual computation, during reexamination phase, is very difficult to measure the reexamination grade scientifically and objectively, reveal the different kinds of competence for each examinee sufficiently. This thesis present our self-developed graduate student reexamination evaluation system. The system exploits some related methods of fuzzy theory to quantize the measurement variables as far as possible. As a result, It can provide a academic basis for advisors to choose the most suitable students, reduce the arbitrary and uncertainty in the process of enrollment further, and enhance justice and clarity as well.Software engineering guideline is implemented throughout the system analysis and design。Moreover, we present a new approach based on the fuzzy theory to process the re-examination data. It achieves the rank of examinees'comprehensive abilities through fuzzy comprehensive judgment for all examinees. Furthermore, It converts an allocation problem to a programming problem which requires the stratification reaches maximum, sets up the integral programming model for this problem, and exploits the classical Hungarian algorithm to achieve the optimal solution. The general profit of the system includes the following aspects:first, it can reduce the admission officers'workload in computation and statistics, and decreases the error rate and randomness for manual input effectively. Second, users can set the parameter for all evaluation items by themselves, and the system preserves some evaluation data setting to be optional in order to enhance the system's flexibility. Third, The system provides scientific and dependable decision support for advisors, so it can advance the efficiency of enrollment greatly. Withal, the system can popularize among the universities and achieve kind of economical benefits.During the whole system designing and implementation, I mainly finished the following work:First, carrying out requirement analysis. Second, transforming the data flow diagram into the software structure diagram, and creating a database model in SQLserver2000 based on summary design of E-R model. Third, applying K-means clustering algorithm to fuzzy treat each ability indicator of examinee, and get the examinee'rank according to the fuzzy comprehensive judgment. Forth, improving Hungarian algorithm, and presenting the optimal allocation method on condition that there are more students than advisors.