04 Fakultät Energie-, Verfahrens- und Biotechnik
Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/5
Browse
2 results
Search Results
Item Open Access Dimensions of proximity and localisation of knowledge-intensive producer services : the case of software services for automotive industry in Stuttgart region(2009) Tözün, Reha; Gaebe, Wolf (Prof. Dr. rer. pol.)This thesis primarily examines the spatiality of knowledge-intensive producer services within the framework of knowledge tacitness and aspatial proximities notions. Additionally, based on an example of a West-European location, it links aspatial proximities and regional structural change discourses and explores the significance of lead firms for regional agglomerations. Globalization operates on the significantly lowered trade barriers and mobility of goods, persons and information. In such an environment, knowledge-production and innovation are critical elements of industrial competitiveness, especially for developed countries in Western Europe. As it has been suggested since late 1980's, innovation has a strong interactive component and the knowledge-creation processes tend to localize geographically. However, such processes are do function over geographical distance as well; therefore one needs to consider the non-geographical influences to understand the dynamics involved. The aspatial proximities, which are categorized by Boschma along cognitive, social, organisational and institutional dimensions, provide an intriguing framework to handle the issues regarding the spatiality of knowledge-producing relations between economic actors. Despite the awareness regarding the role of interaction, theories on regional agglomeration phenomenon tend to be more attentive to the supply-side matters. However, the impulses and incentives from demand side do co-define the direction and extent of knowledge production efforts. With the important roles it assigns to local demand and customers, Porter's industrial clusters concept offers a theoretical background for the research interests of this study. Beside aspatial proximities and industrial clusters, this thesis draws on learning regions, long-waves of technological change and regional structural change literatures for its framework of analysis. For the empirical investigation software services for automotive industry in the Stuttgart region were selected. In-depth interviews with regional experts were adopted for the purposes of data collection. The knowledge-intensive producer services relations fundamentally depend on application- and customer-specific knowledge, which is often change-prone and not market-traded. The study revealed that the well-maintained aspatial proximities allow service providers to have sustained access to such knowledge. It was also found that the creation and maintenance of aspatial proximities is related to the availability of spatial proximity between actors and groups. As such, geographical space has a subtle and indirect effect on knowledge production in that it influences cognitive, social and organisational space through which knowledge is generated and shared. The management of aspatial proximities is also a crucial factor for the path-dependant agglomeration of services activities around existing production locations. The study also found out that the coordination challenges induced by the cognitive dynamics of knowledge-production processes strongly affect the spatial proximity requirements of interactive relations. Cognitive dynamics is defined by the tacitness of knowledge content and the processes characteristics. While tacitness creates the need for face-to-face exchanges, time constraints and interfaces between knowledge-production processes compel actors to have more frequent meetings. Under conditions where highly tacit knowledge content has to be co-produced and shared in short time intervals across numerous processes run by different teams, groups of actors are compelled to engage in more frequent face-to-face interaction in order to avoid knowledge mismatches and miscomprehension. As such, the serviceability of such relations decreases with distance and increasing geographical space between partners escalates transaction costs, which in return impels parties to locate near each other. However, the codifiability of knowledge content and the manageability of processes change these dynamics and spatial proximity becomes a choice rather than a requirement. Provided that their operations contain active technology-oriented and knowledge-production functions, lead firms can act as the nodes of regional knowledge-production networks in clusters. The evidence collected for this study suggests that besides acting as demanding local customers à la Porter, they actively co-develop innovations and enrich knowledge capital of a cluster.Item Open Access Kurzzeit-Prognose von PM10-Konzentrationen in der Luft unter Verwendung maschineller Lernverfahren(2010) Nicklaß, Daniel; Blümel, Wolf Dieter (Prof. Dr.)Galten die mit der Staubbelastung der Luft verbundenen Gefahren und Beeinträchtigungen durch nationale und europäische Umweltrichtlinien Ende der 80er und Anfang der 90er als weitgehend gebannt, so werden nach heutigem Kenntnisstand bereits im Vergleich zu früher relativ geringen Konzentrationen feiner Stäube negative gesundheitliche Auswirkungen zugeschrieben. Erneut in den Blick der Öffentlichkeit geriet die Feinstaubproblematik im Jahre 2005 als nach der Einführung des EU–Grenzwerts für PM10 dieser Grenzwert in deutschen Städten häufig überschritten wurde. Um die zuständigen Behörden und die Bevölkerung über die Luftqualität informieren und gegebenenfallsMaßnahmen zur Vermeidung von Gesundheitsschäden veranlassen zu können, ist es notwendig ein System aufzubauen, das in der Lage ist, PM10-Konzentrationen für den folgenden Tag zu prognostizieren. Hierzu sollen etablierten Prognosemethoden verschiedene innovative maschinelle Lernverfahren (Multilayer-Perzeptron, Support-Vektor-Maschine, Instanzbasiertes Lernen mit den k-nächsten-Nachbarn) gegenübergestellt und die jeweiligen, auf gleicher Datenbasis erzielten Prognoseleistungen miteinander verglichen werden. Da die größte Belastung für die menschliche Gesundheit durch Konzentrationsspitzen mit besonders hohen PM10-Werten ausgehen, liegt hier das besondere Augenmerk der Prognose. In die PM10-Prognose gehen auch meteorologische Vorhersagen ein, die mit großen Unsicherheiten behaftet sind. Durch eine Sensitivitätsanalyse wird untersucht, wie robust die Modelle gegenüber Abweichungen in der Meteorologie sind.