Many real-world vehicle routing problems are dynamic optimization problems, with customer requests arriving over time, requiring a repeated reoptimization. In this paper, we consider a dynamic vehicle routing problem where one additional customer arrives at a beforehand unknown location when the vehicles are already under way. Our objective is to maximize the probability that the additional customer can be integrated into one of the otherwise fixed tours without violating time constraints. This is achieved by letting the vehicles wait at suitable locations during their tours, thus influencing the position of the vehicles at the time when the new customer arrives. For the cases of one and two vehicles, we derive theoretical results about the best waiting strategies. The general problem is shown to be NP-complete. Several deterministic waiting strategies and an evolutionary algorithm to optimize the waiting strategy are proposed and compared empirically. It is demonstrated that a proper waiting strategy can significantly increase the probability of being able to service the additional customer, at the same time reducing the average detour to serve that customer.

Above 90% of the current installed concentrating solar power plants are based on conventional steam-turbine cycles. The operation of steam turbines in these plants is distinctive when compared to traditional base-load power plants. The reason goes back to the intermittent nature of solar power which, in the absence of thermal energy storage or a back-up combustion boiler, forces plant operators to shut down the turbines during night time or at times of low solar radiation. Furthermore, such intermittency often leads to undesirable off-design turbine operating circumstances, either by load variations or changes on live-steam conditions.The present study examines the influence of implementing two operating strategies dealing with steam flow control as a function of incoming solar power for enhancing the thermo-economic performance of a direct steam generation solar tower power plant. The first one consists of a simultaneous high pressure turbine stage- and feed-water preheater bypass. This strategy is used during periods in which the solar radiation is higher than nominal. On these occasions, the plant is capable of generating a larger flow of steam, which allows for an increase in the power production when inserting the additional steam in the turbine bypass. On the other hand, the second operating strategy consists of using an additional feed-water preheater when the power from the field is lower than nominal. In this way, the feed water can reach a higher temperature prior entering the boiler, which is not only beneficial during times of cloud-passages, but also during the start-up process.A dynamic model of a direct steam generation solar tower power plant has been developed following design and operation specifications of an existing reference plant. The two proposed strategies were implemented to the reference model, then a whole year worth simulation was performed for both the reference and the modified models. Lastly, the thermodynamic and economic performance of both systems was measured for the purpose of comparison, by means of using KTH in-house tool DYESOPT. Results show that the implementation of the proposed strategies can enhance the economic viability of the systems by yielding a reduction of 8.7% on the levelized cost of electricity, mainly due to allowing achieving a 12% increase in the net electricity production.

The current demands on Swedish steel industry to produce low quantity batches of specialized products requires research on steel casting processes. There are several physical processes that need be taken into account for this problem to be viewed in full light such as thermal-processes, solidification and fluid dynamics. This work focuses on the fluid-dynamics part; more specifically, the dependence of flow quality within the caster on nozzle and mould geometry. The simulations are carried out using a scale-resolving method, in specific LES (Large Eddy Simulation) which is coupled with a DPM (Discrete Phase Model) to model Argon behaviour. The results of these simulations are presented and validated against physical experiment and data from industrial trials. Conclusions are drawn regarding optimal nozzle types in respect to different mould geometries. The mould eigenfrequencies are shown to exhibit a connection with the casting velocity. This results in so called sweet spots in casting velocity where flow irregularities due to sloshing is minimal. It is shown that the mountain type nozzle is preferable for smaller geometries whilst comparatively larger geometries benefit from a cup type. / FLOWFLEX CC

The work deals with two problems of mass rapid transit system operation: The development of flexible timetables and the realisation of flexible timetables. In both cases, genetic algorithms are used. In the process of (flexible) timetabling in suburban railways, a transport offer perfectly adapted to demand is searched for (temporal and spatial adaptation of demand as well as adaptation of capacity of the trains). After determination of the number of train runs per line and hour and their capacity, optimal departure times have to be found (with a precision of a minute down to 10 s), which fulfil criterias of the passengers (short waiting times) as well as of the operator (small number of vehicles needed). Two different codings for use with genetic algorithms have therefore been developed. They are tested on several case studies of the Dresden suburban railway network, assuming different degrees of flexibilisation. In the process of realising a flexible timetable, transitions between train headways as well as running time and dwell time reserves (margins in the order of a few seconds) are slightly modified in order to coordinate braking and accelerating trains and thereby reduce energy costs of a system of trains. Genetic algorithms can be applied for this problem as well, the proposed methods are tested on several case studies (S-Bahn Berlin, Metro Lille). / Die Arbeit behandelt zwei Probleme der Betriebsplanung von Stadtschnellbahnen: Die Erstellung flexibler Fahrpläne und die Umsetzung flexibler Fahrpläne. In beiden Fällen werden zur Lösung Genetische Algorithmen verwendet. Bei der Ermittlung flexibler Fahrpläne von S-Bahnen wird ein bestmöglich an die Verkehrsnachfrage angepasstes Verkehrsangebot gesucht (zeitlich, räumlich und bezüglich der Kapazität der einzelnen Züge angepasst). Nach stundenfeiner Festlegung der Fahrtenhäufigkeiten und Kapazitäten der einzelnen, sich überlagernden Linien werden deren Abfahrtszeiten gesucht (mit einer Genauigkeit von Minuten bis etwa 10 s), so dass sowohl die Wünsche der Fahrgäste nach gleichmäßigen Zugfolgezeiten als auch Betreiberwünsche (geringe Fahrzeuganzahl) erfüllt werden. Hierzu werden zwei verschiedene Kodierungen für die Verwendung mit Genetischen Algorithmen vorgestellt und das geschaffene Verfahren an verschiedenen Flexibilisierungsszenarien für die S-Bahn Dresden erprobt. Bei der Umsetzung flexibler Fahrpläne, die sich im Bereich weniger Sekunden abspielt, werden Übergänge zwischen Zugfolgezeiten, Fahr- und Haltezeitreserven geringfügig modifiziert, so dass durch bestmögliche Koordination von Anfahr- und Bremsvorgängen eines Systems von Zügen die Energiekosten minimal werden. Methodisch werden wiederum Genetische Algorithmen verwendet, die Erprobung des Verfahrens erfolgt anhand von Linien der S-Bahn Berlin und der Metro in Lille.

The work deals with two problems of mass rapid transit system operation: The development of flexible timetables and the realisation of flexible timetables. In both cases, genetic algorithms are used. In the process of (flexible) timetabling in suburban railways, a transport offer perfectly adapted to demand is searched for (temporal and spatial adaptation of demand as well as adaptation of capacity of the trains). After determination of the number of train runs per line and hour and their capacity, optimal departure times have to be found (with a precision of a minute down to 10 s), which fulfil criterias of the passengers (short waiting times) as well as of the operator (small number of vehicles needed). Two different codings for use with genetic algorithms have therefore been developed. They are tested on several case studies of the Dresden suburban railway network, assuming different degrees of flexibilisation. In the process of realising a flexible timetable, transitions between train headways as well as running time and dwell time reserves (margins in the order of a few seconds) are slightly modified in order to coordinate braking and accelerating trains and thereby reduce energy costs of a system of trains. Genetic algorithms can be applied for this problem as well, the proposed methods are tested on several case studies (S-Bahn Berlin, Metro Lille). / Die Arbeit behandelt zwei Probleme der Betriebsplanung von Stadtschnellbahnen: Die Erstellung flexibler Fahrpläne und die Umsetzung flexibler Fahrpläne. In beiden Fällen werden zur Lösung Genetische Algorithmen verwendet. Bei der Ermittlung flexibler Fahrpläne von S-Bahnen wird ein bestmöglich an die Verkehrsnachfrage angepasstes Verkehrsangebot gesucht (zeitlich, räumlich und bezüglich der Kapazität der einzelnen Züge angepasst). Nach stundenfeiner Festlegung der Fahrtenhäufigkeiten und Kapazitäten der einzelnen, sich überlagernden Linien werden deren Abfahrtszeiten gesucht (mit einer Genauigkeit von Minuten bis etwa 10 s), so dass sowohl die Wünsche der Fahrgäste nach gleichmäßigen Zugfolgezeiten als auch Betreiberwünsche (geringe Fahrzeuganzahl) erfüllt werden. Hierzu werden zwei verschiedene Kodierungen für die Verwendung mit Genetischen Algorithmen vorgestellt und das geschaffene Verfahren an verschiedenen Flexibilisierungsszenarien für die S-Bahn Dresden erprobt. Bei der Umsetzung flexibler Fahrpläne, die sich im Bereich weniger Sekunden abspielt, werden Übergänge zwischen Zugfolgezeiten, Fahr- und Haltezeitreserven geringfügig modifiziert, so dass durch bestmögliche Koordination von Anfahr- und Bremsvorgängen eines Systems von Zügen die Energiekosten minimal werden. Methodisch werden wiederum Genetische Algorithmen verwendet, die Erprobung des Verfahrens erfolgt anhand von Linien der S-Bahn Berlin und der Metro in Lille.