EvoComNet2004 is the first European workshop on the application of evolutionary computation techniques to problems in communication, networks, and connected systems. New communication technologies, the creation of interconnected communication and information networks like the Internet, new types of inter-personal and inter-organisational communication, and the integration and interconnection of production centres and industries are the driving forces on our road towards a connected, networked, society. Workshop papers will be presented orally at the conference and published by Springer as part of EvoWorkshops2004 in the Lecture Notes in Computer Science series:

LNCS 3005, the EvoWorkshops2004 proceedings, is now available online

EvoCOMNET PROGRAMME DETAILS

Workshop PAPERS:

A Scenario-Based Approach to Protocol Design Using Evolutionary Techniques
Sérgio Granato de Araújo, António Carneiro de Mesquita Filho, Aloysio de Castro Pinto Pedroza

Abstract:
An evolutionary approach to design communication protocols from scenario-based specifications is presented. It enables to automatically generate finite-state models of protocol entities from Message Sequence Charts. By converting the Message Sequence Charts into input/output sequences, the problem reduces to evolving finite-state machines with the specified input/output behaviors. The proposed approach does not overgeneralize the entity behavior producing, by construction, minimal, deterministic and completely specified finite-state machines.

Ant Colony Optimization for the Maximum Edge-Disjoint Paths Problem
Maria Blesa, Christian Blum

Abstract:
Given a graph G representing a network topology, and a collection T={(s_1,t_1)...(s_k,t_k)} of pairs of vertices in G representing connection request, the maximum edge-disjoint paths problem is an NP-hard problem which consists in determining the maximum number of pairs in T that can be routed in $G$ by mutually edge-disjoint s_i-t_i paths. We propose an Ant Colony Optimization (ACO) algorithm to solve this problem. ACO algorithms are inspired by the foraging behavior of real ants, whose distributed nature makes them suitable for the application in network environments. Our current version is aimed for the application in static graphs. In comparison to a multi-start greedy approach, our algorithm has advantages especially when speed is an issue.

Using Genetic Programming to Design Broadcasting Algorithms for Manhattan Street Networks
Francesc Comellas, Cristina Dalfó

Abstract:
Broadcasting is the process of disseminating a message from a node of a communication network to all other nodes as quickly as possible. In this paper we consider Manhattan Street Networks (MSNs) which are mesh-structured,
toroidal, directed, regular networks such that locally they resemble the geographical topology of the avenues and streets of Manhattan. With the use of genetic programming we have generated broadcasting algorithms for 2-dimensional and 3-dimensional MSNs.

2-Objective Optimization of Cells Overlap and Geometry with Evolutionary Algorithms
Adel Jedidi, Alexandre Caminada, Gerd Finke

Abstract:
Cellular network design is a very large and complex combinatorial optimization problem. It consists of antenna location and parameters settings. Until now, the design is done using radio quality criteria. Radio coverage, traffic capacity and field overlap are the main factors considered within optimization process to make decisions about network solutions. Nevertheless, such objectives do not lead to an efficient organization of network cells whereas this is a major assessment for radio expert planners. Absence of a clear geometrical structure of network cells prevents experts using many theoretical
concepts on network design. This paper proposes an original model to evaluate the cell shape and a bi-criteria approach using an Evolutionary Algorithm to handle cells overlap and cells geometry as criteria for real-life network optimization.

A Genetic Algorithm for Telecommunication Network Design
Silvana Livramento, Arnaldo Vieira Moura, Flávio Keidi Miyazawa, Mário Massato Harada, Rogério Albertoni Miranda

Abstract:
A Genetic Algorithm (GA) to solve a problem in telecommunication network design is described. The problem is to partition a large urban project area into smaller service sections, which can be controlled by a single standard communication switch. The GA incorporates geometric and topological information from the project area by operating directly with a grid of geographically dispersed demand points. Computational results show this to be a promising technique for partitioning the project area and positioning the control switches. Tests were realized with real instances taken from large areas in the city of Sao Paulo.

A GA/Heuristic Based Hybrid Technique for Routing and Wavelength Assignment in WDM Networks
A Cagatay Talay, Sema Oktug

Abstract:
The routing and wavelength assignment problem which is known to be NP-hard, in all-optical transport networks is considered. The present literature on this topic contains a lot of heuristics. These heuristics, however, have limited
applicability because they have a number of fundamental problems including high time complexity, and lack of scalability with respect to optimal solutions. We propose a hybrid genetic algorithm/heuristic based algorithm. A cost model that incorporates a dependency on link wavelength requirements is adopted. The hybrid algorithm presented uses an object-oriented representation of networks, and incorporates four operators: semi-adaptive path mutation, single-point crossover, reroute, and shift-out. Experimental results of the test networks make clear that, when the network cost depends on heavily wavelength assignment, the proposed GA/Heuristic hybrid approach provides promising results compared to recent wavelength assignment heuristics.

EvoComNet programme committee:
Co-chair: Franz Rothlauf, University of Mannheim <rothlauf@uni-mannheim.de>
Co-chair: George D Smith, University of East Anglia <gds@sys.uea.ac.uk>
Stuart Allen, Cardiff University (UK)
Dave Corne, University of Exeter (UK)
Bryant Julstrom, St Cloud State University (USA)
Joshua Knowles, Universite' Libre de Bruxelles (Belgium)
Geoff McKeown, UEA Norwich (UK)
Martin Oates, University of Reading (UK)
Günther Raidl, Vienna University of Technology (Austria)
Franz Rothlauf, Mannheim (Germany)
George D. Smith, UEA Norwich (UK)
Giovanni Squillero, Politecnico di Torino (Italy)
Andrew Tuson, City University, London (UK)

EvoWorkshops chairs:
Günther Raidl, Vienna University of Technology <raidl@ads.tuwien.ac.at>
Stefano Cagnoni, Universita' di Parma <cagnoni@ce.unipr.it>
Local chair : Ernesto Costa, University of Coimbra <ernesto@dei.uc.pt>

Communication systems includes telecommunications; mobile, satellite, optical, and voice communications; personal communication systems; switching and routing; transmission systems; access systems; communication systems simulation; station and antenna design; information and speech processing; intrusion detection; error control coding; compression and cryptography; propagation and channel modelling, etc.

Networks includes networks and graph problems; unconstrained and constrained network design problems; network flows; network and computer security; Internet problems; electrical, power, and data networks; computer networks; location and link design; reliability and failure; corporate network design; location placement; network physical and software architecture; network hardware and software technologies; mobility networks and protocols; operations, maintenance, and management; signalling and control; active networks; network services and applications, etc.

Connected systems includes network problems in non-technical environments; manufacturing, distribution and logistics networks; supply and disposal networks; inter-personal communication; communication analysis; inter-organisational communication; supply chains; information networks, etc.

All aspects of evolutionary computation methods, as well as other nature-inspired and non-evolutionary design and optimization techniques, are welcome.

The reviewing process is double blind. Authors should remove their names from submitted papers, and should take reasonable care that their identity is disguised. References to own work can be included in the paper, but should be referred to in the third person.

The EvoWorkshops2004 proceedings will be
published by Spinger as part of their
Lecture Notes in Computer Science series.