JOURNAL TITLE - MONTH YEAR 1 Multicast Heuristic Approaches on Multi- Layer Wireless Network Mauro Tropea 1 , Amilcare Francesco Santamaria 2 DEIS Department, University of Calabria via P. Bucci cubo 42/c, Arcavacata di Rende 87036 – Cosenza - Italy 1 mtropea@deis.unical.it; 2 afsantamaria @deis.unical.it Abstract Multicast services are increased in exponential manner in these last few years and in particular in the field of multimedia services. Multicast can reduce resources allocation and enhance network performances in terms of QoS, especially in a wireless platforms where the bandwidth is a precious resource. The problem of multicast routing can be reduced to the problem of finding a spanning tree capable to distribute network flow among multicast sources and destinations. It has been established that determining an optimal multicast tree for a static multicast group can be modeled as the Steiner Tree problem in networking, this problem has been proofed to be a NP-complete. Hence the necessity of using scalable algorithms in scalable networks composed of multi layered platforms. Moreover, in this work a QoS multi-constraint multicast problem has been addressed. In this paper a comparison between two meta- heuristic algorithms is presented in order to show the scalability introduced by these types of algorithms that are able of finding sub-optimal solutions. These meta-heuristics are based on Genetic Algorithm and Simulated Annealing mechanism. A simulated campaigns between two proposed algorithm has been addressed. Keywords Scalable Algorithm; Multicast routing; heterogeneous platforms; Genetic Algorithm; Simulated Annealing Introduction This The request of multicast services are increased in exponential manner in these last few years and in particular in the field of multimedia services. Multicast can reduce resources allocation and enhance network performance in terms of QoS and resources allocation. Key factor of the multicast is the capacity to reduce packets number that flows on the network. Only necessary packets are sent by the source of a generic multicast group, after that the network will provide to send spread towards all destinations data flow. In the multicast routing two entities work together with the main goal to distribute data among all nodes that belong to a multicast groups, these entities are the multicast protocol and the multicast algorithm, this last one can be completely disconnected from protocols, in this case the protocol has the task to trigger algorithm when needed. Use integrated or independent algorithm depends of network data to be distributed or from application type, moreover, several approaches exist to implements multicast such as centralized or distributed, shared or not, static or dynamic and so on. With rapidly growth of hardware technologies and the rapid evolution of Internet even more kind of applications was born, that require QoS constraints. Multicast routes its packets following a multicast tree that is searched or built in order to reach each destination. The problem of multicast routing can be reduced to the problem of finding a tree spanning the source and all destinations that belong to the multicast group. It has been established that determining an optimal multicast tree for a static multicast group can be modeled as the Steiner tree problem in networks, which is proofed to be a NP- complete problem. Hence the necessity of using scalable algorithms in scalable networks composed of multi layered platforms [1][2], in particular a multi- constraints QoS multicast problem is faced. In order to face the scalable issues also mechanisms of Call Admission Control can be previewed for better exploiting wireless resources [3]. Many mechanisms exist that try to make the multicast algorithms scalable. They are heuristic search whose purpose is to find a sub-optimal solution. Some mimics the process of natural evolution as Genetic Algorithms; others are based on the collective behavior of decentralized, self- organized systems, known as Swarm Intelligence; others are inspired on the metallurgic process called annealing that is a technique involving heating and controlled cooling of a material to increase the size of its crystals and reduce their defects. In this paper a comparison between two meta-heuristic algorithms is