Abstract. Many algorithms have been proposed for scheduling real-time tasks to guarantee the timing requirements of such a system. Each of those algorithms is based on a decision parameter with which the algorithm determines the order of the tasks to be executed. Among those parameters, deadline and laxity are the most important ones which have been employed in the Earliest Deadline First (EDF) and the Least Laxity First (LLF) algorithms respectively [1]. Furthermore, to include the implicit constraints imposed by real world, such as uncertainty and lack of complete knowledge about the environment, dynamicity in the world, bounded validity time of information and other resource constraints, a few fuzzy approaches have been proposed [2, 3]. It was shown that the fuzzy approaches outperform the non-fuzzy approaches. The main purpose of this paper is to study and compare the behavior of a real-time system when the fuzzy scheduler is based on deadline and when it is based on laxity. The obtained results show that deadline is a better parameter to be considered in fuzzy real-time scheduling. Keywords: Real-time Scheduling, Fuzzy Logic, EDF, LLF I. INTRODUCTION Scheduling real time systems involves allocation of resources and CPU-time to tasks in such a way that certain performance requirements are met. In real-time systems scheduling plays a more critical role than non-real-time systems because in these systems having the right answer too late is as bad as not having it at all [1]. Such a system must react to the requests within a fixed amount of time which is called deadline. Real-time tasks can be classified as periodic or aperiodic. A periodic task is a kind of task that occurs at regular intervals, and aperiodic task occurs unpredictably. The length of the time interval between the arrivals of two consecutive requests in a periodic task is called period. There are a plenty of real-time scheduling algorithms that are proposed in the literature. Each of these algorithms bases its decision on certain parameter while attempting to schedule tasks to satisfy their time requirements. Some algorithms use parameters that are determined statically such as the Rate Monotonic algorithm that uses the request interval of each task as its priority [4, 7]. Others use parameters that are calculated at run time. Laxity and deadline are among those parameters that are the most considered. Laxity says the task execution must begin within a certain amount of time while deadline implies the time instant at which its execution must be completed [2, 5, 6]. The main purpose of this paper is to study and compare the behavior of a real-time system when the fuzzy scheduler is based on deadline and when it is based on laxity. The rest of the paper is organized as follows. In Part II we will give an introduction to fuzzy inference process. Part III covers the algorithms and part IV presents the experimental results. Conclusion and future works are debated in Sections V. II. 2. FUZZY INFERENCE SYSTEMS A general fuzzy inference system consists of three parts. A crisp input is fuzzified by input membership functions and processed by a fuzzy logic interpretation of a set of fuzzy rules. This is followed by the defuzzification stage resulting in a crisp output [8]. There are a number of different ways to implement the fuzzy inference engine. Among the very first such proposed techniques is that due to Mamdani [10], which describes the inference engine in terms of a fuzzy relation matrix and uses the compositional rule of inference to arrive at the output fuzzy set for a given input fuzzy set. The output fuzzy set is subsequently defuzzified to arrive at a crisp control action. Fig 1. Block diagram of fuzzy inference process The inference engine is based on a collection of logic rules in the form of IF-THEN statements, where the IF part is called the "antecedent" and the THEN part is called the Deadline vs. Laxity as a Decision Parameter in Fuzzy Real-Time Scheduling Mojtaba Sabeghi, Koen Bertels Mahmoud Naghibzadeh {sabeghi, k.l.m.bertels}@ce.et.tudelft.nl naghib@um.ac.ir Delft University of Technology Ferdowsi University of Mashhad Delft, the Netherlands Mashhad, Iran