Genetic algorithm based intelligent test paper generation

In recent years, due to the rapid development of technology, it has become possible to store and use the data effectively in an electronic environment as well as in all areas of life. The innovations provided by technology in this era, where electronic media and paperless solutions are becoming more and more widespread, make our lives easier every day. The exams that are applied for the measurement and evaluation studies which are part of the education system can be prepared in the electronic environment and can be made independent of time and place, thus offering easy access to large communities. The test questions prepared for the exams can also be stored and used in the electronic environment together with their various properties. Along with the use of electronic question banks in the education and training, the need for the preparation of test pages in the desired characteristics emerges. In order to generate intelligent test pages, different methods and algorithms are used instead of human being, so that high quality and effective test pages can be generated at the desired level while saving time. There are studies in which traditional algorithms and methods are used for generating a test page on the desired test page from a question bank. However, the application of these methods often increases the solution time, can cause situations such as the test page being the same as the one generated in the past and may reduce the yield for test page production. Heuristic methods are often used as an alternative for problems that are difficult to solve by mathematical or traditional methods or that have a long solution time. These problems can be overcome by the heuristic optimization methods used in the artificial intelligence and optimum results can be obtained according to the desired criteria even in a question bank having limited questions. With the heuristic optimization methods, the solution can be brought to the best possible by going through the randomly selected questions without any aim of producing the best solution. The optimum result can be found for the solution by evaluating the many required criteria. In the thesis study, a problem was solved by using an heuristic approach to the problem of generating test pages with multiple criterias and at the same time, it was aimed to reduce human labor force and time loss. The genetic algorithm adapted to the problem of the test page has been provided in a fast and efficient manner to generate the test page in the desired criteria and specifications. Web-based application software has been implemented in order to make the test page generation in this work easier for users. The genetic algorithm used in the study is compared with the standard genetic algorithm results and it is seen that the genetic algorithm used in the study gave better results.