Content uploaded by Pierre Kuonen
Author content
All content in this area was uploaded by Pierre Kuonen on Apr 02, 2015
Content may be subject to copyright.
Genetic approach to radio network optimization for mobile systems
Abstract
Size and complexity of future UMTS radio networks make their
planning very difficult. This paper focuses on the radio coverage
problem, that is, to cover a maximum surface of a given geographical
region at an optimal cost. This combinatorial optimization problem is
solved with a bioinspired genetic algorithm. A first prototype runs in
parallel on a network of workstations. The obtained results are shown
and discussed
... The coverage optimisation problem [8], [9] is also known in the literature as the Antenna Positioning Problem (APP) [2], [10], [11], [12], [13], the Radio Network Design (RND) [4], [6], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], the Antenna Placement Problem (APP) [14], [24], [25], [26], [27], [28], [29], [30], the Automatic Cell Planning (ACP) [31], [32], [33] or the Radio Network Optimisation (RNO) [34], [35], [36], [37], [38]. It is also called the Base Station Placement (BSP) [39], [40], [41], [42], [43], [44], [45], [46] or finally the Base Station Location (BSL) [47], [48], [49], [50], [51], [52] . ...
... It was born from the works done in the STORMS project (2) [37]. It is based on a graph formulation introduced by Calegari [16], [34], [35], [36], [37] that seeks to treat a more canonical and simplified model of the APP. It regroups only the fundamental objectives of this task such as the amount of covered area and the number of used antennas. ...
... Unlike the ARNO model, there is no distinction between the type of points and no complex propagation loss matrix is used to simulate the wave propagation. Simple isotropic or directive wave propagation models are used [24], where cells coverage can be computed and returned by an ad-hoc function [35]. ...
... Las otras etapas de esta metodología incluyen asignación de frecuencias, canales dentro una frecuencia, etc. Pero está claro que sin una cobertura de señal el resto de los problemas planteados en el diseño de una red no pueden surgir. Según [11] el problema de RND pertenece a la clase NP-Completa, lo que permite que el mismo sea abordado mediante el uso de metaheuríticas. Una de las características más importantes de las metaheurísiticas es su generalidad para ser aplicadas a cualquier problema [9], y en base a las lecturas realizadas sobre comunicaciones [6], [14] y [15] el modelo matemático es el que determina el tipo de servicio que se va a implementar. ...
... El estado del arte desarrollado en [10] establece como referencia de investigación un algoritmo genérico para la resolución del problema de RND desarrollado por [11] basado en teoría de grafos. A partir de este trabajo se llevaron adelante diversas investigaciones utilizando AGs para resolver el problema de RND, como los que se presentan en [22], [23] y [25] hasta la aplicación de modelos paralelos [9], [20], [21] y [24]. ...
... El estudio pretende ofrecer una base de referencia confiable sobre un amplio espectro de algoritmos y medidas precisas de comparación de la eficiencia, confiabilidad y rapidez de las diferentes técnicas aplicadas a la resolución del RND. En este trabajo se destaca a [11] quién utilizó una AG aplicado a comunicaciones móviles a mediados de los 90 y por el desarrollo de un entorno de trabajo conocido como STORMS (Software Tools for Optimization of Resource in Mobile Systems). ...
La difusión de servicios inalámbricos de comunicación (telefonía, internet, etc) es cada vez mayor en la actualidad. El costo de los equipamientos necesarios para brindar el servicio con calidad adecuada es elevado. La selección de un conjunto de puntos geográficos que permitan una cobertura óptima de una señal de radio frecuencia minimizando uso de recursos es fundamental. A esta tarea se la denomina diseño de la red de radio y es un problema NP – duro de optimización, por lo tanto, es factible de ser tratado con metaheurísiticas. Las metaheurísiticas son métodos que integran procedimientos de mejora local y estrategias de alto nivel para realizar una búsqueda robusta en el espacio del problema.
El presente trabajo propone analizar y evaluar la capacidad de tres algoritmos genéticos canónicos de encontrar una solución aceptable mediante una función objetivo que combina el grado de cobertura del terreno y el uso eficiente de recursos en diferentes escenarios poblacionales. Los resultados obtenidos son promisorios.
... In the latter works, more flexible and efficient optimisation methods [Ama01,Ama02,Ama03], concerning the advanced statistical search heuristics like tabu search [Glo93,Glo97,Cap02], simulated annealing [Kir83,Kir84,Oso96,Tut99] or genetic algorithms [Cal97,Gui97,Rei99,Toe04] necessary for preparation of new network parameters measured in seconds necessary for preparation of new network parameters measured in seconds necessary for preparation of new network parameters measured in seconds necessary for preparation of new network parameters measured in seconds ---almost on almost on almost on almost on----line). Achievement of this goal is possible by the synthesis of a novel line). ...
... If an optimisation task can be modelled on the basis of the continuous linear object, it can be solved almost perfectly and very quickly using the Linear Integer Programming (LIP) methods. It is important to say that linear optimisation problems are very well worked out in theory 14 As the inherent structure of the EAs enables a parallel implementation on many machines [Cal97,Gui97,Toe04], this concept seems to be a great opportunity to speed up the calculations and to enhance the area of their potential applications. Unfortunately, this approach is beyond the possibilities of the author of this thesis as a quite expensive and rather complicated solution. ...
This is my (Adam Sobczyk) PhD thesis (dissertation) - mostly focused on automatic, very fast methods ("on-line") of the WCDMA/3G/UMTS networks capacity and coverage optimization, according to the traffic spatial distribution.
... One of the problems in the development of mobile telecommunications network is the selection of proper placement of BTS locations, so as to reach customers optimally with adequate traffic services. To overcome this problem, some optimization techniques placement of base stations has been carried out, ie mobile network optimization with genetic approaches [4] and the application of genetic algorithms for optimization of radio network coverage [5]. ...
... In this work, using the sectoral antenna in 120 0 to form the coverage area of a circle. Correction cell radius is expressed: r = d x 27.. ….. (4) Where r is the radius correction cells, whereas d is the radius of a cell planning results. ...
The increasing number and the distribution expansion of mobile subscribers need the cellular network to expand its service coverage, especially in urban areas. It encourages cellular operators to continually extend its cellular coverage area by constructing new towers of base transceiver station (BTS) in certain areas. BTS development that continues to grow in the limited urban areas causing environmental clutter. So it requires rearranging the presence of the tower. This paper proposes a novel method to minimise the number of existing cellular towers that able to cover an urban area of Malang city. The Fuzzy C-Means and Particle Swarm Optimization methods are used to optimise the problem. Briefly, the proposed method is to make a GSM standard-based cellular network planning to obtain the ideal number of cells, and then to optimise the BTS's composition for finding the best tower with the widest regional coverage. In this study, there are two types of cellular providers based on the frequency spectrum used. The result showed that there are 48 of 154 existing towers selected. The result also revealed that the new BTS's composition covers 75.39% of the total area for the frequency bandwidth of 10 MHz, while this new composition covers 64.49% of the total area for the frequency bandwidth of 7.5 MHz..
... Si une solution a déjà été testée et si elle ne convient pas, elle sera appelée ainsi "tabou" afin que l'algorithme ne l'envisage plus. Ce type d'approche d'optimisation par heuristiques [20][21][22][23][24][25][26] et plus particulièrement des algorithmes génétiques [27][28][29] est également appliqué. ...
... En Mendes et al, (2009) se realiza una revisión general de diferentes metaheurísiticas resolviendo el problema de RND y utilizando la función objetivo propuesta en y Calegari (b) et al, (1997). El estudio pretende ofrecer una base de referencia confiable sobre un amplio espectro de algoritmos y medidas precisas de comparación de la eficiencia, confiabilidad y rapidez de las diferentes técnicas aplicadas a la resolución del RND. ...
En el presente trabajo se aplica una serie de versiones del algoritmo genético no convencional denominado Cross generational elitist selection Heterogeneous recombination Cataclysmic mutation algorithm (CHC ) para resolver el problema de diseño de red de radio (RND). Se utiliza un conjunto de algoritmos genéticos para realizar una comparativa de rendimiento de los algoritmos propuestos. Se emplea una función objetivo basada en la eficiencia de iluminación de la señal. Se utiliza la variabilidad genética de la población como parámetro de convergencia y detección de incesto y se propone el uso de la variabilidad del mejor individuo como mecanismo de sacudida. Esto permite generar poblaciones dinámicas conforme a las soluciones más promisorias generando diferentes espacios de búsqueda. Los resultados obtenidos por los algoritmos propuestos son satisfactorios.
The precise placement of antennas is essential to ensure effective coverage, service quality, and network capacity in wireless communications, particularly given the exponential growth of mobile connectivity. The antenna positioning problem (APP) has evolved from theoretical approaches to practical solutions employing advanced algorithms, such as evolutionary algorithms. This study focuses on developing innovative web tools harnessing genetic algorithms to optimize antenna positioning, starting from propagation loss calculations. To achieve this, seven empirical models were reviewed and integrated into an antenna positioning web tool. Results demonstrate that, with minimal configuration and careful model selection, a detailed analysis of antenna positioning in any area is feasible. The tool was developed using Java 17 and TypeScript 5.1.6, utilizing the JMetal framework to apply genetic algorithms, and features a React-based web interface facilitating application integration. For future research, consideration is given to implementing a server capable of analyzing the environment based on specific area selection, thereby enhancing the precision and objectivity of antenna positioning analysis.
In this paper we study the position planning of cellular antennas. Informally, we try to find the best positions of antennas by having the maximum coverage and the minimum cost. Several methods were used to solve this problem. Since we have two main objectives, metaheuristics, such as genetic algorithms, simulated annealing, particle swarm optimization,. .. were used. Actually, artificial intelligence, and more particularly deep learning are compared to other methods in this context. Our work is about applying reinforcement learning on this problem. In this vein, we tried to tune some reinforcement learning methods to solve this problem. We found that the SARSA agent with the EpsGreedy QPolicy is the best option for these kinds of problems.
Radio network planning is a core problem in cellular networks. It includes coverage, capacity and parameter planning. This paper investigates the Antenna Positioning Problem (APP) which is a main task in cellular networks planning. The aim is to find a trade-off between maximizing coverage and minimizing costs. APP is the task of selecting a subset of potential locations where installing the base stations to cover the entire area. In theory, the APP is NP-hard. To solve it in practice, we propose a new meta-heuristic called Evolutionary Iterated Local Search that merges the local search method and some evolutionary operations of crossover and mutation. The proposed method is implemented and evaluated on realistic, synthetic and random instances of the problem of different sizes. The numerical results and the comparison with the state-of-the-art show that the proposed method succeeds in finding good results for the considered problem.
One of the issues telecommunication companies must face when deploying a mobile phone network is the selection of good locations for base transceiver stations (BTSs). The problem comes down to guarantying a maximum percentage of radio coverage in a region with a minimum number of BTSs. The best subset of BTS locations must thus be selected among a set of user-provided potential ones to achieve this goal. This paper introduces a way to solve this problem. First, a graph-based modelling of the BTS placement problem is defined. Second, an overview of genetic algorithms is given, and a parallel approach is proposed. Finally, the results obtained with the Paragene program are discussed.
The ParFlow method permits to simulate outdoor wave propagation in urban environment. It compares with the so-called Lattice Boltzman Model (LBM), which describes a physical system in terms of the motion of fictitious microscopic particles over a lattice. One of the objectives of the European project STORMS1 (Software Tools for the Optimization of Resources in Mobile Systems) is to develop a software tool to be used for design and planning of the UMTS network2. This paper gives an overview of the ParFlow method, and reports the design and the implementation of ParFlow++, a parallel object-oriented software for outdoor wave propagation prediction in urban environment. ParFlow++ is being developed in C++, and the resulting code is primarily targeted at MIMD-DM platforms. Although the development of ParFlow++ is still in progress in the STORMS project, a prototype version is available, that compiles and runs on a network of Unix workstations, as well as on the Cray T3D. Experiments achieved with this prototype software lead to promising results, which are discussed in this paper.
Radio Network Optimization With Maximum Independent Set Search
- B Chamaret
- S Josselin
- P Kuonen
- M Pizarosso
- S.-B Manzanedo
- S Ubeda
- D Wagner