Traffic Engineering for Hard Real Time Multicast Applications Using Genetic Algorithm

In this paper, the problem of finding a near-optimal routing solution for a set of real time applications with multicast traffic flows in a meshed core network is studied. Such kind of network is typically a core network of automobile, airplane, or industry control networks. We will introduce our routing planning methodology based on a genetic algorithm, and will focus on the model with Earliest Deadline First (EDF) scheduler. The properties and schedulability conditions of EDF scheduler are also analyzed in this paper. We will show how hard real time QoS can be guaranteed through proper routing and delay bounds allocation. At last, based on the calculation result, the performance of our methodology will also be discussed. Although this work originates from the analysis of some industrial core networks, the methodology introduced in this paper is also suitable for solving similar problems in other networks.