There are various technologies used for reducing fuel consumption of automobiles. Hybrid electric vehicles is one approach that has been used, which can reduce fuel consumption by 10-30% compared to conventional vehicles. In this master thesis the minimization of fuel consumption of a power-split type HEV along a given route is considered, where the vehicle speed has been assumed to be known a priori. This minimization was made by first deriving a model of the HEV powertrain, followed by creating a Dynamical programming based program for finding the optimal distribution of torques. The performance was evaluated through the commercial software GT-Suite. The resulting control from the Dynamic program could follow the reference speed in many situations. However the battery state-of-charge calculated in the Dynamic program did not update properly, resulting in a depleted battery in some cases. The model derived could follow the dynamics of the vehicle, but there are some parts which could be improved. One of them is the dynamical model of the rotational speed for the engine <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Comega_%7Be%7D" />.  The Dynamic program works for finding the controller, and can be modified to work with improved state-equations. / Det finns olika sätt att minska bränsleförbrukningen hos bilar, men ett sätt som använts är el-hybrider. Dessa kan minska bränsleförbrukningen med 10-30% jämfört med konventionella bilar. I det här examensarbetet undersöks optimering av bränsleförbrukning för en el-hybrid, där hastigheten antas vara känd i förväg. Optimeringen skedde genom att först härleda en modell för drivlinan, och därefter skapades ett Dynamisk programerings baserat program för att hitta den optimal kombinationen av moment. Bränsleförbrukning och prestanda jämfördes genom programvaran GT-Suite. Dynamiska programmeringen gav lovande resultat som följde referenshastigheten i många fall. Däremot uppdaterades inte laddningen för batteriet lika bra, vilket ledde till att batteriet i vissa fall blev urladdat. Modellen som härleddes visade i många fall liknande respons som GT-Suite, men viss förbättring kan ske. En utav dessa förbättringar är rotationsekvationen för bränslemotorn, <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Comega_%7Be%7D" />. Den Dynamiska programmeringen som skapades fungerade, och kan modifieras för förbättrade tillståndsekvationer

Optimal control theory has a long history and broad applications. Motivated by the goal of obtaining insights through unification and taking advantage of the abundant capability to generate data, this thesis introduces some suboptimal schemes via abstract dynamic programming models. As our first contribution, we consider deterministic infinite horizon optimal control problems with nonnegative stage costs. We draw inspiration from the learning model predictive control scheme designed for continuous dynamics and iterative tasks, and propose a rollout algorithm that relies on sampled data generated by some base policy. The proposed algorithm is based on value and policy iteration ideas. It applies to deterministic problems with arbitrary state and control spaces, and arbitrary dynamics. It admits extensions to problems with trajectory constraints, and a multiagent structure. In addition, abstract dynamic programming models are used to analyze $\lambda$-policy iteration with randomization algorithms. In particular, we consider contractive models with infinite policies. We show that well-posedness of the $\lambda$-operator plays a central role in the algorithm. The operator is known to be well-posed for problems with finite states, but our analysis shows that it is also well-defined for the contractive models with infinite states. Similarly, the algorithm we analyze is known to converge for problems with finite policies, but we identify the conditions required to guarantee convergence with probability one when the policy space is infinite regardless of the number of states. Guided by the analysis, we exemplify a data-driven approximated implementation of the algorithm for estimation of optimal costs of constrained linear and nonlinear control problems. Numerical results indicate the potentials of this method in practice. / Teorin om optimal reglering har en lång historia och breda tillämpningsområden.I denna avhandling, som motiveras av att få insikter genom att förena och dra nyttaav den goda möjligheten att generera data, introduceras några suboptimala systemvia abstrakta modeller för dynamisk programmering.I vårt första bidrag betraktar vi ett deterministiskt optimalt regleringsproblemmed oändlig horisont och icke-negativa stegkostnader. Vi hämtar inspiration frånmodellprediktiv reglering med inlärning, som är utformad för system med kontinuerligdynamik och iterativa uppgifter, och föreslår en utrullningsalgoritm som bygger påsamplade data som genereras av en viss baspolicy. Den föreslagna algoritmen byggerpå idéer om värde- och policyiteration. Den är tillämpningsbar för deterministiskaproblem med godtyckliga tillstånds- och kontrollrum samt för system med godtyckligdynamik. Slutligen kan den utvidgas till problem med trajektoriebegränsningar ochen struktur med flera agenter.Dessutom används abstrakta modeller för dynamisk programmering för attanalysera lambdapolicyiteration med randomiseringsalgoritmer. Vi betraktar merspecifikt kontraktiva modeller med oändliga strategier. Vi visar att lambdaoperatorns välbestämdhet spelar en central roll i algoritmen. Det är känt att operatorn ärväldefinierad för problem med ändliga tillstånd, men vår analys visar att den ocksåär väldefinierad för de studerade kontraktiva modellerna med oändliga tillstånd.På samma sätt är det känt att den algoritm vi analyserar konvergerar för problemmed ändliga strategier, men vi identifierar de villkor som krävs för att garanterakonvergens med sannolikhet ett när policyrummet är oändligt, oberoende av antalettillstånd. Med hjälp av analysen exemplifierar vi en datadriven approximativ implementering av algoritmen för uppskattning av optimala kostnader för begränsadelinjära och icke-linjära regleringsproblem. Numeriska resultat visar på potentialen iatt använda denna metod i praktiken. / <p>QC 20211129</p>

Information availability plays a fundamental role in decision-making for business operations. The present dissertation aims to develop frameworks and algorithms in order to guide a decision-maker in environments with limited information. In particular, in the first part, we study the fundamental problem of designing optimal auctions while relaxing the widely used assumption of common prior. We are able to characterize (near-)optimal mechanisms and associated performance. In the second part of the dissertation, we focus on data-driven pricing in the low sample regime. More precisely, we study the fundamental problem of a seller pricing a product based on historical information consisting of one sample of the willingness-to-pay distribution. By drawing connection with the statistical theory of reliability, we propose a novel approach, using dynamic programming, to characterize near-optimal data-driven pricing algorithms and their performance. In the last part of the dissertation, we delve into the detailed practical operations of the online display advertising marketplace from an information structure perspective. In particular, we analyze the tactical role of intermediaries within this marketplace and their impact on the value chain. In turn, we make the case that under some market conditions, there is a potential for Pareto improvement by adjusting the role of these intermediaries.

Indiana University-Purdue University Indianapolis (IUPUI) / Due to the increasing demand from mobile devices for the real-time response of cloud computing services, federated edge learning (FEL) emerges as a new computing paradigm, which utilizes edge devices to achieve efficient machine learning while protecting their data privacy. Implementing efficient FEL suffers from the challenges of devices' limited computing and communication resources, as well as unevenly distributed datasets, which inspires several existing research focusing on device selection to optimize time consumption and data diversity. However, these studies fail to consider the energy consumption of edge devices given their limited power supply, which can seriously affect the cost-efficiency of FEL with unexpected device dropouts. To fill this gap, we propose a device selection model capturing both energy consumption and data diversity optimization, under the constraints of time consumption and training data amount. Then we solve the optimization problem by reformulating the original model and designing a novel algorithm, named E2DS, to reduce the time complexity greatly. By comparing with two classical FEL schemes, we validate the superiority of our proposed device selection mechanism for FEL with extensive experimental results. Furthermore, for each device in a real FEL environment, it is the fact that multiple tasks will occupy the CPU at the same time, so the frequency of the CPU used for training fluctuates all the time, which may lead to large errors in computing energy consumption. To solve this problem, we deploy reinforcement learning to learn the frequency so as to approach real value. And compared to increasing data diversity, we consider a more direct way to improve the convergence speed using loss values. Then we formulate the optimization problem that minimizes the energy consumption and maximizes the loss values to select the appropriate set of devices. After reformulating the problem, we design a new algorithm FCE2DS as the solution to have better performance on convergence speed and accuracy. Finally, we compare the performance of this proposed scheme with the previous scheme and the traditional scheme to verify the improvement of the proposed scheme in multiple aspects.

A stochastic shortest path problem is an undiscounted infinite-horizon Markov decision process with an absorbing and costree target state, where the objective is to reach the target state while optimizing total expected cost. In almost all cases, the objective in solving a stochastic shortest path problem is to minimize the total expected cost to reach the target state. But in probabilistic model checking, it is also useful to solve a problem where the objective is to maximize the expected cost to reach the target state. This thesis considers the maximum-time stochastic shortest path problem, which is a special case of the maximum-cost stochastic shortest path problem where actions have unit cost. The contribution is an efficient approach to computing high-quality bounds on the optimal solution for this problem. The bounds are useful in themselves, but can also be used by other algorithms to accelerate search for an optimal solution.

No description available.

In recent years, healthcare management has become fertile ground for the scheduling theory community. In addition to an extensive academic literature on this subject, there has also been a proliferation of healthcare scheduling software companies in the marketplace. Typical scheduling systems use rule-based analytics that give schedulers advisory information from programmable heuristics such as the Bailey-Welch rule cite{B,BW}, which recommends overbooking early in the day to fill-in potential no-shows later on. We propose a dynamic programming problem formulation to the scheduling problem that maximizes revenue. We formulate the problem and discuss the effectiveness of 3 different algorithms that solve the problem. We find that the 3rd algorithm, which has smallest amount of nodes in the decision tree, has an upper bound given by the Bell numbers. We then present an alternative problem formulation that includes stochastic appointment lengths and no shows.

<P> For over six decades, the theory and design of radar systems have been dominated by probability theory and statistics, information theory, signal processing and control. However, the similar encoding-decoding property that exists between the visual brain and radar has been sadly overlooked in all radar systems. This thesis lays down the foundation of a new generation of radar systems, namely cognitive radar, that was described in a 2006 seminal paper by Haykin. Four essential elements of cognitive radar are Bayesian filtering in the receiver, dynamic programming in the transmitter, memory, and global feedback to facilitate computational intelligence. All these elements excluding the memory compose a well known property of mammalian cortex, the perception-action cycle. As such, the cognitive radar that has only this cycle is named as the basic cognitive radar (BCR). For t racking applications, t his thesis presents the underlying theory of BCR, with emphasis being placed on the cubature Kalman filter to approximate the Bayesian filter in the receiver, dynamic optimization for transmit-waveform selection in the transmitter, and global feedback embodying the transmitter , the radar environment, and the receiver all under one overall feedback loop. </p> <p> Built on the knowledge learnt from the BCR, this thesis expands the basic perception-action cycle to encompass three more properties of human cognition , that is, memory, attention, and intelligence. Specifically, the provision for memory includes the three essential elements, i. e. , the perceptual memory, executive memory, and coordinating perception-action memory that couples the first two memories. Provision of the three memories adds an advanced version of cognitive radar, namely the nested cognitive radar (NCR) in light of the nesting of three memories in the perception-action cycle. </p> <p> In this thesis, extensive computer simulations are also conducted to demonstrate the ability of this new radar concept over a conventional radar structure. Three important scenarios of tracking applications are considered, they are (a), linear target tracking; (b), falling object tracking; and (c), high-dimensional target tracking with continuous-discrete model. All simulation results confirm that cognitive radar outperforms the conventional radar systems significantly. </p> <p> In conducting the simulations, an interesting phenomenon is also observed, which is named the chattering effect. The underlying physics and mathematical model of this effect are discussed. For the purpose of studying the behaviour of cognitive radar in disturbance, demonstrative experiments are further conducted. Simulation results indicate the superiority of NCR over BCR and t he conventional radar in low, moderate and even strong disturbance. </p> / Thesis / Doctor of Philosophy (PhD)

No description available.

No description available.