Cloud computing provides the illusion of a seamless, infinite resource pool with flexibleon-demand accessibility. However, behind this illusion there are thousands ofservers and peta-bytes of storage, running tens of thousands of applications accessedby millions of users. The management of such systems is non-trivial because theyface elastic demand, have heterogeneous resources, must fulfill diverse managementobjectives, and are vast in scale.Autonomic computing techniques can be used to tackle the complex problem ofresource management in cloud data centers by introducing self-managing elementsknown as autonomic managers. Each autonomic manager should be capable of managingitself while simultaneously contributing to the fulfillment of high level systemwideobjectives. A wide range of approaches and mechanisms can be used to defineand design these autonomic managers as well as to organize them and coordinate theiractions in order to achieve specific goals.This thesis investigates autonomic approaches for cloud resource management thataim to optimize the cloud infrastructure layer with respect to various high level objectives.The resource management problem is formulated as a problem of optimizationwith respect to one or more management objectives such as cost, profitability, or datacenter utilization, as well as performance concerns such as response time, quality ofservice, and rejection rates. The aim of the reported investigations is to address theproblems of cost-efficient elastic resource provisioning, unified management of cloudresources, and scalability in cloud resource management. This is achieved by introducingthree new concepts in capacity management: the Repacking, Holistic, and Peerto Peer approaches.

No description available.

Purpose:To investigate how managers work with capacity dimensioning and what the main challenges are in order to balance efficiency and responsiveness in the continuous operations, as well as investigate what patterns and trends that can be identified within the capacity dimensioning approach. Methodology:A multiple case study was conducted including 14 manufacturing companies. Empirical data was collected through semi-structured interviews and used to explain the phenomenon of capacity dimensioning. Differences and similarities in the way companies approach capacity dimensioning was investigated though a cross-case analysis. The research is of exploratory and inductive character. Findings:A general process for capacity dimensioning has been established and affecting aspects and challenges has been identified. Potential trends and relationships have been investigated for the capacity dimensioning approach, with a potential connection between flexibility and investment strategy with introduction period in human resources. Further the capacity strategies tend to vary depending on alternative capacity sources. Theoretical implications:Information is provided for how capacity dimensioning is done at companies today, connections are strong to adjacent theories as S&OP but with more detail in the area of setting the capacity level. Managerial implications:The capacity dimensioning does not have a solution that suits all companies, but communication and alignment in the supply chain should not be underestimated for successful capacity dimensioning. Research delimitations:The research is conducted on manufacturing companies active in Sweden, other countries or continents of the world could generate other results because of different culture and laws. Further, service companies could also have provided other results.

Parallel computing on heterogeneous workstation clusters has proved to be a very efficient use of available resources, increasing their overall utilization. However, for it to be a viable alternative to expensive, dedicated parallel machines, a number of key issues need to be resolved. One of the major challenges of heterogeneous computing is coping with the inherent heterogeneity of the system, with the availability of workstations from different vendors of varying processing speeds and capabilities. The existence of multiple jobs and users further complicates the task. The time taken for a parallel job is constrained by the time taken by the slowest or the most heavily loaded workstation. Therefore, load sharing of parallel computations is imperative in ensuring good overall utilization of the system. Since load sharing is essentially independent of the particular parallel job being run, the development of program independent, automatic, scheduling and load sharing strategies have become vital to the efficient use of the heterogeneous cluster. This thesis discusses various prior approaches to load sharing, examines a new strategy developed for heterogeneous workstations, and evaluates its performance. / Graduation date: 1996

Revenue management is a concept aimed to maximize capacity utilization and through that maximize revenues. It originated in the airline industry in the 70’s and due to its effectiveness  quickly spread to other sectors of the service industry. Today it is used in several industries like hotels, television and radio broadcasters, and energy transition companies to name a few. Since revenue management was developed in and for the service industry, most studies on revenue management are done on the service industry, creating a rather large research cap. Recently this concept has spread to the manufacturing industry as well. Despite this, there is very limited research done on revenue management in the manufacturing industry. Therefore, this paper’s aim is to partially filling this research gap by studying capacity management and pricing strategies (two mechanisms of revenue management), and how they have been shaped when implemented in a manufacturing company. This paper was done with a case study done on a multinational manufacturing company, who recently implemented revenue management. Interviews were conducted with people in key positions with good insight to the usage of revenue management in this company. Some of the most important result was that in this manufacturing company it is not possible to nest capacity on a customer segment level. However, in this company nesting was done on a market level instead. Also the pricing strategy differed between the service industry theory and this company. Instead of having a dynamic price that changed the total price up or down to change demand, this company had more of a fixed total price, and instead added more features to the product, decreasing the profit margin. The conclusion was drawn that the industry characteristics of the manufacturing industry have forced a rather large modification of revenue management. However, since this was a qualitative case study, no generalizing conclusions for the entire manufacturing industry can be drawn.

published_or_final_version / Electrical and Electronic Engineering / Master / Master of Philosophy

abstract: This research develops heuristics to manage both mandatory and optional network capacity reductions to better serve the network flows. The main application discussed relates to transportation networks, and flow cost relates to travel cost of users of the network. Temporary mandatory capacity reductions are required by maintenance activities. The objective of managing maintenance activities and the attendant temporary network capacity reductions is to schedule the required segment closures so that all maintenance work can be completed on time, and the total flow cost over the maintenance period is minimized for different types of flows. The goal of optional network capacity reduction is to selectively reduce the capacity of some links to improve the overall efficiency of user-optimized flows, where each traveler takes the route that minimizes the traveler’s trip cost. In this dissertation, both managing mandatory and optional network capacity reductions are addressed with the consideration of network-wide flow diversions due to changed link capacities. This research first investigates the maintenance scheduling in transportation networks with service vehicles (e.g., truck fleets and passenger transport fleets), where these vehicles are assumed to take the system-optimized routes that minimize the total travel cost of the fleet. This problem is solved with the randomized fixed-and-optimize heuristic developed. This research also investigates the maintenance scheduling in networks with multi-modal traffic that consists of (1) regular human-driven cars with user-optimized routing and (2) self-driving vehicles with system-optimized routing. An iterative mixed flow assignment algorithm is developed to obtain the multi-modal traffic assignment resulting from a maintenance schedule. The genetic algorithm with multi-point crossover is applied to obtain a good schedule. Based on the Braess’ paradox that removing some links may alleviate the congestion of user-optimized flows, this research generalizes the Braess’ paradox to reduce the capacity of selected links to improve the efficiency of the resultant user-optimized flows. A heuristic is developed to identify links to reduce capacity, and the corresponding capacity reduction amounts, to get more efficient total flows. Experiments on real networks demonstrate the generalized Braess’ paradox exists in reality, and the heuristic developed solves real-world test cases even when commercial solvers fail. / Dissertation/Thesis / Doctoral Dissertation Industrial Engineering 2017

No description available.

No description available.

A well-functioning transport system that can ensure efficient distribution is an important part of a mass-producing company. This is also the case for Holmen Paper Braviken, a paper mill with a widespread market base. Both today and historically, a large market share has been found in Europe, where a large proportion of the paper produced has been transported by truck and boat. In recent times, the demand for container transports has become higher as a result of sales in increasingly widespread markets. It is a trend that is forecasted to continue and to ensure that increased volumes henceforth still can be transported, Holmen Paper Braviken's current container transport solution has become interesting to investigate. In this study, the current process for container transports is examined as an outset and basis for identifying a possible solution that in the long term can facilitate a suitable development. It is clear that the current process for container transport is standardized with few adaptable possibilities but is well-functioning under the present conditions. However, hard endeavors are needed in specific periods due to high variations in volumes, which suggests a need to be flexible in adjustments. Furthermore, a container warehouse was identified as a possible development for the existing process. A container warehouse can lead to higher flexibility in transport activities by enabling the possibility to allocate and plan transports and resources to a greater extent. Since loading of containers is made possible when the undertakings are low, which can ease coming periods where there is a spike in demand. Therefore, loading operations can be spread out better over time and resources can more easily be redistributed based on the existing conditions of demand.   However, the introduction of a container warehouse is currently not seen as a feasible change since it involves significant investments and risks. If, on the other hand, the development on the market continues according to the current pattern, said investment will become a more and more reasonable solution. Since increasingly higher demands will be placed on flexible container transports, which would warrant a container warehouse to a higher degree. / Ett välfungerande transportsystem som kan säkerställa en effektiv distribution är en viktig del i ett massproducerande företag. Så är även fallet för Holmen Paper Braviken, ett pappersbruk med en spridd marknadsbas. Både idag och historiskt har en stor marknadsandel befunnit sig inom Europa där en stor andel av det producerade pappret transporterats med lastbil och båt. Under senare tid har dock efterfrågan på containertransporter blivit högre till följd av försäljning på allt mer spridda marknader. Det är en trend som prognostiserats fortsätta och för att säkerställa att ökande volymer fortsatt kan hanteras har Holmen Paper Bravikens nuvarande transportlösning för container blivit intressant att undersöka. Som utgångsläge i den här studien undersöktes den nuvarande processen för containertransporter för att identifiera en eventuell lösning som på sikt kan underlätta för den aktuella händelseutvecklingen. Det framgår att nuvarande process för containertransporter är standardiserad och väl fungerande under dagens förutsättningar men med få anpassningsmöjligheter. I perioder förekommer det ansträngningar på nuvarande system, främst baserat på höga variationer i volymer, vilket ställer krav på en flexibel anpassningsförmåga. Vidare identifierades ett containerlager som en tänkbar utveckling till den befintliga processen. Ett containerlager kan leda till en högre flexibilitet i transportaktiviteter genom att fördela om resurser och i högre utsträckning möjliggöra långsiktig planering av transporter. Eftersom förlastning av containrar är möjligt vid låg beläggning kan man underlätta för kommande tidsperioder med hög beläggning. Det innebär att beläggningen kan jämnas ut och resurser kan i perioder enklare fördelas om efter givet tillstånd i beläggning. Införandet av ett containerlager ses i dagsläget däremot inte som en genomförbar förändring eftersom det innebär betydande investeringar och risker. Om utvecklingen på marknaden däremot fortsätter enligt nuvarande mönster kommer sagda investering bli mer och mer aktuell. Då kommer allt högre krav ställas på flexibla containertransporter där ett införande av ett containerlager skulle kunna bli berättigat.