Stefan Türk

Network migration using ant colony optimization

This paper describes the basic concepts of the network migration problem, as well as its cost optimal solution with an ant colony algorithm. Generally there are many possible approaches to calculate network migration strategies. One of those is to consider the migration as a Shortest Path Problem, which can be solved using meta heuristics. In this paper it will be analyzed how ant colony optimization can be used to calculate migration sequences with minimized costs. Results of our investigation are information about the speed of convergence of ant colony algorithms towards an optimal cost value. The used resource utilization to cover increasing network demands will be presented also.

Approaches for the migration of optical backbone networks towards Carrier Ethernet

This paper describes migration strategies in optical networks. It is expected that the conversion of current transport networks (IP/MPLS over SDH) into more packet oriented (IP/MPLS over Ethernet) technologies will result in significant cost savings for network operators. The state space of a migrating network can be modeled as a directed graph. To find the optimal migration path within this state oriented graph, several approaches, such as the Incremental Migration and the All-Period Migration can be used. Current research results regarding cost development and resource utilization in layer 1 and layer 3 are presented in this work. The effects of several planning techniques within such uncertain environment, like routing and grooming, on the total cost development are discussed, too.

Node Degree based Improved Hop Count Weighted Centroid Localization Algorithm

Hop-count based weighted centroid localization is a simple and straightforward localization algorithm, which uses anchors with known positions and the hop count to these anchors to estimate the real position of nodes. Especially in sensor networks, where energy restrictions prevent more complex algorithms, this fast and simple algorithm can be used. Unfortunately the localization error of the algorithm can hinder the practical usage. In this paper we will improve the weighted centroid algorithm for hop count based localization by adding the node degree on the paths to the referenced anchors into the weights. After an analysis to obtain theoretically optimal coefficients we will show by means of simulation that for longer hop counts to the anchors and areas with different node degrees the proposed ND-WCL algorithm outperforms the known hop count based weighted centroid localization algorithm.

Optimization of energy efficient network migration using harmony search

In this paper we describe the basic network migration problem in backbone networks to move from an existing to a new technology. Furthermore we use a generic harmony search algorithm to optimize the solution in terms of energy efficiency and costs. Harmony search is a probabilistic meta-heuristic which has been successfully adapted to many optimization problems. We analyze how harmony search can be used to calculate migration sequences with minimized energy consumption and financial costs. Results of parameter studies for the heuristic will be shown to evaluate the method. The achieved resource utilization to cover increasing network demands and the point of introduction within a certain time interval will be presented also.

Network Migration Optimization Using Genetic Algorithms

This paper introduces the problem of communication network migration for backbone networks. Heuristic solutions for this problem can be determined by the application of genetic algorithms to the problem. A description of the system model is presented, as well as the used algorithmic approaches and optimization results. Our main goal is the optimization of migration costs, by respecting increasing demands over the migration period, while device costs per bit are decreasing. We will present Crowded DPGA as best found GA to solve the network migration problem.

Improving network migration optimization utilizing memetic algorithms

In this paper we describe the basics of the network migration problem, as well as the solution of the problem using meta heuristics. The class of heuristics described and applied in this paper is called memetic algorithms. Those are explained and results regarding network migration are presented. To evaluate potential algorithm candidates a generic TSP has been implemented and tested. The mapping of the network migration problem to memetic algorithms and to the algorithm library GALib is explained as well. Our main goal is the optimization of network migration costs, by respecting increasing demands over the migration period, while device costs per bit are decreasing. The mapping of the specified problem to the heuristic variants is also of major interest.

New IEEE 802.16-2009 compliant traffic shaping algorithms for WiMAX networks

WiMAX based on the IEEE 802.16 standard allows the deployment of cellular mobile broadband access networks delivering all services through the Internet Protocol. An essential feature of such networks it the Quality of Service (QoS) support including the traffic shaping. Based on detailed analyses of the IEEE 802.16-2009 QoS concept, we propose two traffic shaping algorithms. The first one is standard conform at the expense of high memory consumption. The second approximate algorithm has been developed for memory crucial system. We present parameter studies obtained by a detailed ns-2 WiMAX simulation environment. The simulation results show that both algorithms can be used as a base for QoS-aware WiMAX scheduling algorithms.

Particle Swarm Optimization of network migration planning

Traffic demand of backbone networks will increase rapidly in the next years, while equipment costs per bit are shrinking. To cope with this development the migration of those backbones towards Carrier Ethernet is investigated. In this paper we discuss a Particle Swarm Optimization (PSO) heuristic that can be used to enhance the overall migration process. We will introduce the basic migration problem, as well as system models for a possible migration case. Improved heuristics for basic PSO and their application to network migration will be presented. Results regarding algorithm performance and optimized migration cost will be delivered in the paper.

Strategic migration optimization of urban access networks using meta-heuristics

In this paper the concept of network migration and its relation to access networks is introduced. A typical urban migration case (from VDSL to GPON) is studied to provide cost results under given boundary conditions. We investigate the optimization of a migration process over a long period of time using a segmentation of huge urban scenarios into clusters. Although those clusters can be separately migrated and optimized, an evaluation, which cluster shall be optimized in a certain time step is necessary. This strategic optimization can be performed heuristically. The application of those methods to exemplary scenarios is described in the paper. The necessary sequencing of these steps and our achieved results in terms of costs and algorithm performance are explained. The paper provides methods to cope with network changes and increasing demands via usage of planning heuristics.

Optimization of 2d virtual node coordinates in anchor-free localization algorithms for geographic routing in ad-hoc networks

Many localization algorithms require that at least some nodes in a network know their correct real position. Anchor-free localization algorithms manage to assign to all nodes virtual coordinates (VC) without knowledge of real location information. This is mostly done by node cooperation. These algorithms provide VCs for the usage of geographic routing in environments where real coordinate localization is impossible due to cost, time, energy, space or line of sight restrictions. This paper proposes improvements in generating VCs for wireless ad-hoc networks without location information. We present algorithms for improving the preparatory phase, which is used to obtain initial VCs as well as for the working phase, which adapts to network changes and node movement. Using simulations for different node densities and network sizes, we compare different approaches and show which algorithms outperform others.