A Hierarchical Genetic Algorithm for Topology Optimization and Traffic Management in ATM Networks Utilizing Bandwidth and Throughput
Alaa F. Sheta, Malik Shehadeh Braik, Mohammad Salamah, Saudi Arabia, Prince Abdullah, Bin Ghazi · 2010
Abstract: Th eAsy nc rou a f Md ( ) tw k is expected to become a backbone network for high speed digital multimedia services in distributed environments. This paper explores an optimization based hierarchical approach using Genetic Algorithms (GAs) for installing an ATM network with optimal structure and traffic flow. A Hierarchical Genetic Algorithm (HGA) is used to solve such an optimization problem. HGA approach can improve the performance of conventional GA, though it consumes more computation time. Thus, HGA approach is capable of reducing the overall delay while increasing the bits transmitted over the network. This will definitely improve the network performance and meets the requirements in multimedia service environments. The preliminary results indicate that HGA based ATM network design can be very efficacious. Keywords: G en tic A lgo rithms ,TM N tw ok F w Management. 1. Introduction There is a rapidly demand for telecommunication networks to function efficaciously despite obstacles such as disabling portions of the network, limiting the links capacities, high cabling costs, and so on [1]. Thus, networks must constantly able to maintain an acceptable level of performance. Despite this need, the problem of dynamic redesign of functioning networks still received a little attention. Today, with the intensifying role of using the Internet and networking technologies, there have been immense advances in using the telecommunications services in the present information society. Asynchronous Transfer Mode (ATM) is one of the most promising networking technologies. ATM was rated to reduce the complexity of the network and improve the flexibility of traffic performance which has good ability for transmitting many kinds of information, and carry traffic over all kinds of networks. Furthermore, ATM has been increasingly used in telecommunications systems for high speed multimedia services [2], [3]. Recently, control flow and management are core concepts of the optimal structure of ATM networks; hence, flow control and management have drawn much attention by the researchers in telecommunication field. In addition to that the nodes in the ATM network design must be linked in an economical way to handle expected traffic and capacity constraints [1], [4]. As the models for the design of ATM networks are quite complex, and one of the limiting factors is the requirement of expensive exchange based equipments, th erfo , G n ti cA lg oit m s(GAs) w ud v h problems of traffic assignment and topological design of local area networks [5], [6], flow optimization in survivable ATM networks [7], and flow assignment in ATM networks [8]–[10]. However, many programming models have been developed which deals with telecommunication network planning [11]. In this context, Evolutionary Algorithms (EAs) have been frequently applied to telecommunication services in the last years [9], [12]. This paper handles an optimal ATM network structure (i.e. topology and link capacities) using Hierarchical Genetic Algorithms (HGA) [13], [14]. The objectives of the network planning are: designing the network structure to carry out the estimated traffic, and minimizing the cost of network [15],[16]. Consequently, we are dealing with flow assignment and links capacities as complex optimization problems using GAs in two levels, the first GA level deals with selection the optimal network structure, and the second GA level reduces the overall delay while maintain the throughput, this cycle will continues until terminated by some condition. The remainder of the paper is organized as follows. In Section 2 we explain the details of GA levels; each alternative level of HGA is described extensively. The case