Abstract
In the environment of RTOS, in order to determine its success, the system not only has to ensure the accuracy of task performance, but also finishes in time. Although EDF(Earliest Deadline First) on the uniprocessor computers has had operation scheduling algorithm as the best scheduling method, multiprocessor instant scheduling system, because of the complication, so far the best scheduling algorithm is still not available. Schedulers are available to simulate multiprocessor instant system, but proper suggestions might not be presented for users to comprehend better solutions while desired scheduling is impossible. Therefore, in this dissertation, the adaptive scheduler will be presented. Not only users will be informed whether it is acceptable to proceed scheduling by using simulation result, but also it will present comparative suggestions according to the related situation changes within the system under the circumstance of unsolved scheduling. In addition, the scheduler can assist users to solve scheduling.