Changes in big companies resulting in new organizational structures and cost cutting are pushing more and more of the knowledge and information handling to sub-units in a multi-national structure. For big knowledge-intensive companies that act in the global market place, internal information handling is becoming a challenge. The study and reflections are based on experiences from Ericsson, a knowledge-intensive global telecommunication company. This company delivers complex cross-functional products (solutions) and has a decentralized organization. It faces the cost of managing distributed product information and the challenge to gather relevant information in the sales departments. One can easily characterize the company’s complex and unique product offerings as having multiple dependencies. The solutions are composed by building blocks, i.e. different sub products, delivered by different product units. The different sub products suffer from limitations in how they can be combined into solutions. This study addresses the information gaps in a decentralized organization regarding this specific issue. It focuses on identifying vital information without driving cost and requiring organizational changes. Stakeholder identification was done from a value chain perspective. The type of information that would give the most profitable solutions was identified during group sessions and individual interviews. An asymmetric compatibility matrix (ACM) was developed to fit the purpose of keeping low maintenance cost and without requiring organizational changes. The ACM was applied and process maturity improvements were evaluated with the use of the Process Enterprise and Maturity Model. The Ericsson specific study shows that the use of an ACM for product compatibility information makes it possible to define information responsibility that is sustainable over time. Thereby the maintenance cost for this information can be brought down to a minimum. Furthermore, the study shows that the effort of gathering information for the sales organizations to provide customer solutions can be reduced by the use of an ACM offering generic compatibility information. Users of the ACM would be able to re-use its information and focus on customer specific sales and deployment issues rather than re-do what others already have done. Cost of sales as well as business risks would thereby likely decrease, affecting the bottom line positively. Furthermore new business opportunities are assumed to be addressed better since relevant generic information is made available up front. Other positive expected benefits are prevention of network malfunctions and increased customer satisfaction. From this study it can be concluded that an ACM can be a powerful tool for gathering cross-functional product information in large decentralized organizations at a low cost, without any organizational changes, and with a high process maturity. Further research would be needed if one would consider validating the general applicability of the ACM in other processes.

The Vehicle Routing Problem has been around for more than 50 years and has been of major interest to the operations research community. The VRP pose a complex problem with major benefits for the industry. In every supply chain transportation occurs between customers and suppliers. In this thesis, we analyze the use of a multiple pheromone trial in using Ant Systems to solve the VRP. The goal is to find a reasonable solution for data environments of derivatives of the basic VRP. An adaptive object model approach is followed to allow for additional constraints and customizable cost functions. A parallel method is used to improve speed and traversing the solution space. The Ant System is applied to the local search operations as well as the data objects. The Tabu Search method is used in the local search part of the solution. The study succeeds in allowing for all of the key performance indicators, i.e. efficiency, effectiveness, alignment, agility and integration for an IT system, where the traditional research on a VRP algorithm only focuses on the first two. / Thesis (PhD)--University of Pretoria, 2010. / Industrial and Systems Engineering / unrestricted

[pt] Um algoritmo de alocação de freqüência eficiente deve considerar aspectos de propagação e detalhes de equipamento de modo a minimizar a relação a sinal interferência, e conseqüentemente, aumentar a eficiência do espectro. Essa dissertação tem o objetivo de propor uma melhor solução para alocação de freqüência nas rede TDMA e GSM, levando em consideração os aspectos citados. A abordagem utilizada é diferente das técnicas de alocação de freqüências usuais, que são baseadas na distância de reuso. O algoritmo proposto utiliza uma matriz que representa os aspectos de propagação (matriz de isolação) e uma matriz que representa os detalhes do equipamento (matriz de compatibilidade). A matriz de isolação mostra a relação sinal interferência que cada célula tem com as outras. Um estudo de como coletar os dados (seleção da rota no interior do setor) e de como trabalhar com os dados coletados foram realizados, objetivando-se ter matriz que melhor caracterizasse a rede utilizada. A matriz de compatibilidade mostra a distância em freqüência no qual se devem alocar os canais em uma mesma célula, células adjacentes e células co-sites. Os valores de espaçamento entre freqüências, sugeridos nesta dissertação, são valores práticos e dependem das técnicas utilizadas no planejamento e do equipamento utilizado em questão. O problema de alocação foi modelado com um problema inteiro. Uma heurística (divide and conquer) foi utilizada para particionar a rede em menores partes, em que o algoritmo branch and bound foi implementado. O algoritmo foi testado em duas redes com sessenta e trezentos e quarenta e oito setores respectivamente. As redes foram construída levando em consideração os dados topográficos e morfológicos da cidade de Brasília. / [en] An efficient frequency assignment algorithm should consider both propagation conditions and equipment details, in order to minimize the carrier-to-interference ratio (C/I) and, as a result, to increase spectrum efficiency. This dissertation proposes a way to obtain a better solution for frequency assignment in TDMA and GSM networks, taking into account propagation conditions and equipment details. The proposed approach is different from usual frequency assignment techniques based on reuse distance. A frequency assignment algorithm proposed uses one matrix that represents the propagation features (isolation matrix) and another which represents the equipment details (compatibility matrix). The isolation matrix shows the signal to interference ratio between any two sectors in the whole network. A survey to define how to collect the data ( selection of the route inside the sector) and how to work with data collected was developed, aiming to obtain the best matrix representation of the network. The compatibility matrix shows the frequency distance between channels in the same sector, adjacent sector and co-site sectors. The frequency space values used in this matrix are practical values, which depend upon the technique used on the planning and on the utilized hardware. The frequency assignment was formulated as an integer problem. An heuristic (divide and conquer) was used to split the network into smaller parts where an exact algorithm branch and bound was implemented. The algorithm was tested on two networks with 60 and 348 sectors respectively. The networks were built having the topographical and morphological data from Brasilia city. / [es] Un algoritmo eficiente de asignación de frecuencia debe considerar aspectos de propagación y detalles del equipo de modo tal que se minimize la relación señal-interferencia, y onsecuentemente, aumente la eficiencia del espectro. Esta disertación tiene como objetivo proponer una solución para la asignación de frecuencia en las redes TDMA y GSM, considerando los aspectos anteriormente citados. El enfoque utilizado difiere de las técnicas de asignación de frecuencias usuales, baseadas en la distancia de reuso. EL algoritmo propuesto utiliza una matriz que representa los aspectos de propagación (matriz de aislamiento) y una matriz que representa los detalles del equipo (matriz de compatibilidade). La matriz de aislamiento muestra la relación señal-interferencia que cada célula tiene con las otras. Se realiza un estudio de como recolectar los datos (selección de la ruta en el interior del sector) y de como trabajar con los datos recolectados, con el objetivo de obtener la matriz que mejor caracteriza la red utilizada. La matriz de compatibilidad muestra la distancia en frecuencia en la cual se deben alocar los canales en una misma célula, células adyacentes y células co-sites. Los valores de espacio entre frecuencias, sugeridos en esta disertación, son valores prácticos y dependen de las técnicas utilizadas en el planificación y del equipo utilizado. El problema de asignación fue modelado como un problema entero. Fue utilizada una heurística (divide and conquer) para particionar la red en menores partes, donde se implementó el algoritmo branch and bound. EL algoritmo fue probado en dos redes con sesenta y trezentos y quarenta y oito sectores respectivamente. Las redes fueron construidas considerando los datos topográficos y morfológicos de la ciudad de Brasilia.