A fuzzy scheduling model based on multi-characteristics of tasks in embedded soft real-time Systems oriented to communication field

Abstract

In the task running environment oriented to communication field applications, the characteristcs of a task are commonly undeterministic and unpredictable due to the unsteadiness of communication networks. It is necessary to apply fuzzy concept and theory into the embedded soft real-time task scheduling field. In this paper, we presented a fuzzy scheduling model based on multi-characteristics of a task in soft real-time systems oriented to communication field. In this model, we adopted fuzzy scheduling algorithm in which a ready task's scheduling turn is decided by the fuzzy inference result of its fuzzified characteristics such as criticality and deadline distance. In the fuzzy inference, we presented two kinds of inference mode:the nearest similarity principle and Mamdani principle. The test result shows that our presented fuzzy scheduling model has less deadline miss ratioes to the more important tasks than pure EDF algorithm when the test system is overloaded.