An airship with a moving gondola is investigated with the goal of achieving a large pitch angle and of minimizing the total energy consumption required to goal position. The airship in this study is equipped with different actuation tools such as a moving gondola and a vectoring thrust, which can be used for various flight modes. The efficiency of the actuation methods employed is studied and compared in various flight scenarios, based on the airship’s ability to reach the desired position while consuming the least amount of energy. The nonlinear dynamic model is derived using Newton-Euler equations. Backstepping and incremental nonlinear dynamic inversion (INDI) controllers are designed to track a desired trajectory by controlling the position of the gondola and the thrust. The dynamic models are then implemented and simulated in the Matlab/Simulink to evaluate the effectiveness of the controllers in different environmental conditions. The simulation results show the effectiveness of the controllers used, and a larger pitch angle of -89o, can be reached thanks to the movement of the gondola to the front of the airship on the curved keel. The airship prototype was used for the experimental test to evaluate the pitch tracking performance of each of the controllers. The experimental results show that the prototype used can generate a -90o pitch angle, thereby improving manoeuvrability and allowing for rapid changes in altitude. The energy model is developed to evaluate and compare the energy required by the airship for ascent, cruise, and descent flights, using different actuation methods. The effectiveness of the composite control strategy is demonstrated by completing the flight mission with the least amount of energy consumed. An optimization method is then developed to find the optimum design configuration to reduce the cost function, based on energy consumption, of the different flight scenarios while always respecting the design constraints. The Heuristic technique is used to obtain the optimal flight trajectory based on the platform’s ability to complete the desired mission while minimising energy consumption. The results show that for pitch tracking, the vectored thrust has a rapid response, and the required thrust is high. Therefore, this configuration requires more energy than the moving gondola control configuration, in all cases studied. The composite configuration is found to be the most efficient method for completing the flight trajectory with the least amount of energy. The total energy consumption of the entire flight is reduced by about 17% by using the optimization algorithm to select the best actuation method for each flight mode.

Thesis (S.M.)--Massachusetts Institute of Technology, Computation for Design and Optimization Program, 2011. / Page 118 blank. Cataloged from PDF version of thesis. / Includes bibliographical references (p. 112-117). / The Bayesian approach to inference problems provides a systematic way of updating prior knowledge with data. A likelihood function involving a forward model of the problem is used to incorporate data into a posterior distribution. The standard method of sampling this distribution is Markov chain Monte Carlo which can become inefficient in high dimensions, wasting many evaluations of the likelihood function. In many applications the likelihood function involves the solution of a partial differential equation so the large number of evaluations required by Markov chain Monte Carlo can quickly become computationally intractable. This work aims to reduce the computational cost of sampling the posterior by introducing a multiscale framework for inference problems involving elliptic forward problems. Through the construction of a low dimensional prior on a coarse scale and the use of iterative conditioning technique the scales are decouples and efficient inference can proceed. This work considers nonlinear mappings from a fine scale to a coarse scale based on the Multiscale Finite Element Method. Permeability characterization is the primary focus but a discussion of other applications is also provided. After some theoretical justification, several test problems are shown that demonstrate the efficiency of the multiscale framework. / by Matthew David Parno. / S.M.

Thesis: S.M., Massachusetts Institute of Technology, Computation for Design and Optimization Program, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 61-65). / Dynamical systems are subjected to various random external forcings that complicate theie control. In order to achieve optimal performance, these systems need to continually adapt to external disturbances in real time. This capability is provided by feedback based control strategies that derive an optimal control from the current state of the system. Model Predictive Control(MPC) is one such feedback-based technique. This thesis explores the application of a stochastic version of MPC to a reservoir system. The reservoir system is designed to maximize the revenue generated from the hydroelectricity while simultaneously obeying several exogenous constraints. An ensemble based version of the stochastic MPC technique is studied and applied to the reservoir to determine the optimal water release strategies. Further analysis is performed to understand the sensitivity of different parameters in the MPC technique. / by Reetik Kumar Sahu. / S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Computation for Design and Optimization Program, 2009. / This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 77-78). / Modeling of multiphase flow in fractured media plays an integral role in management and performance prediction of oil and gas reserves. Geological characterization and nmultiphase flow simulations in fractured media are challenging for several reasons, such as uncertainty in fracture location, complexity in fracture geometry. dynamic nature of fractures etc. There is a need for complex sinmulation models that resolve the flow dynamics along fractures and the interaction with the porous matrix. The unstructured finite volume model provides a tool for the numerical simulation of multiphase flow (inmmiscible and incompressible two-phase flow) in two-dimensional fractured media. We use a finite volume formulation, which is locally imass conservative and it allows the use of fully unstructured grids to represent the coimplex geometry of the fracture networks. Fractures are represented as objects of lower diniensionality than that of the domain (in this case, ID objects in a 2D domain). The model permits fine-scale simulation of multiphase transport through fractured media. The non-Fickian transport resulting due to the presence of heterogeneity (as fractures or inhomogeneous permeability distribution) is captured by the traditional advection-diffusion equation using a highly discretized system. Today. many macroscopic flow models are being developed which account for the non-Fickian. non-local flow more accurately and efficiently with less computation. The finite volume simulator niodel described in this thesis will be instrumental as a tool to train and validate the macroscopic flow models which account for anomialous transport behavior. / (cont.) We illustrate the performance of this simulator on several synthetic cases with different fracture geometries and conclude the model effectively captures the miiultiphase fluid flow pattern in fractured media. / by Reena Bajaj. / S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Computation for Design and Optimization Program, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 77-79). / Hydrogen fuel is considered one of the major emerging renewable substitutes for fossil fuel. A crucial factor as to whether hydrogen will be successful depends on its cost as a substitute. Recently, there has been a growing interest in investigating the feasibility from a supply chain point of view. This thesis intends to provide a comprehensive study of the feasibility of hydrogen as a transportation fuel from a supply chain point of view. The aim is to discover the most efficient, sustainable, and ultimately most cost-effective strategy to meet future transportation demand scenarios. This includes optimizing costs over production, compression, storage, distribution, and dispensation. Moreover, through the decision support model developed in this thesis, insights regarding strategic approaches for hydrogen supply chain infrastructure development are developed, such as the tradeoff between centralized and distributed production. A case study in Spain is presented to illustrate the supply chain. Different models are proposed to estimate electricity generation capacity, hydrogen production scheme, transportation topology, distribution methods and future demands. Six different scenarios, based on different production scheme, future demand level, are tested. Results from the case study indicate that it is more cost effective to transmit electricity from wind farms to locations close to demand sites to do centralized production. In terms of transportation, liquid gas truck is the preferred mode of transportation from production sites to local demand regions. The model can adapt be extended and adapted to consider other configurations of the supply chain network. / by Li Liang. / S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Computation for Design and Optimization Program, 2006. / Includes bibliographical references (p. 93-94). / One of the steps in stowage planning is crane split and sequencing, which determines the order of container discharging and loading jobs quay cranes (QCs) perform so that the completion time (or makespan) of ship operation is minimized. The vessel's load profile, number of bays and number of allocated QCs are known to port-planners hours before its arrival, and these are input parameters to the problem. The problem is modeled as a large-scale linear IP where the planning horizon is discretized into time intervals and at most one QC can be assigned to a bay at any period. We introduce clearance constraints, which prevent adjacent QCs from being positioned too close to one another, and yard congestion constraints, which prevent yard storage locations from being overly accessed at any time. This makes the model relevant in an industrial setting. We examine the case only a single ship arrives at port, and the case where multiple ships berth at different times in the planning horizon. The berth time of each ship and number of ships arriving is known. The problem is difficult to solve without any special technique applied. For the single-ship problem, a heuristic approach, which produces high-quality solutions, is developed. / (cont.) A branch-and-price method re-formulates the problem into a set-covering form with huge number of variables; standard variable branching provides optimal solutions very efficiently. For the multiple-ship problem, a solution strategy is developed combining Lagrangian relaxation, branch-and-price and heuristics. After relaxing the yard congestion constraints, the problem decomposes into smaller sub-problems, each involving one ship; the sub-problems are then re-formulated into a column generation form and solved using branch-and-price to obtain Lagrangian solutions and lower-bound values. Lagrangian multipliers are iteratively updated using the sub-gradient method. A primal heuristic detects and eliminates infeasibilities in the Lagrangian solutions which then become an upper bound to the optimal objective. Once the duality gap is sufficiently reduced, the sub-gradient routine is terminated. The availability of efficient commercial modeling software such as OPL Studio and CPLEX allows for larger instances of the problem to be tackled than previously possible. / by Shawn Choo. / S.M.

Thesis: S.M., Massachusetts Institute of Technology, School of Engineering, Center for Computational Engineering, Computation for Design and Optimization Program, 2015. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 75-76). / Today, organizations are generating large volumes of data. However, the challenge of extracting valuable information from the data has been a large and long-standing problem. Here, we address the problem of quantifying risks and detecting fraud in heterogeneous financial big data. Great financial losses are pressuring institutions to devise innovative solutions for risk and fraud detection. Current approaches in government suffer from issues such as high false positive rates and low adaptability to the continuous evolution of newer fraud. In this thesis, we propose an open and extensible framework called "Situational Awarness FrAamework for RIsk ranking" (SAFARI). SAFARI aims to quantify and rank risk with unlabeled, complex data in the financial world. The framework integrates and analyzes different perspectives of financial data, and extends risk scores for decision makers. SAFARI also utilizes machine learning techniques to learn from examined cases to improve the calculation of risks and adapt to the changing behavior of fraudulent activities. The work includes designing, implementing, testing, extending and evaluating the proposed framework. In the overpayment detection scenario, results show SAFARI can effectively find overpayments with low false positive rates. Furthermore, SAFARI can be extended to assist decision making in a variety of environment thanks to its general applicability. / by Rongsha Li. / S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Computation for Design and Optimization Program, 2008. / Includes bibliographical references (p. 75-76). / In this thesis, we study the Projective Re-Normalization method (PRM) for semidefinite programming feasibility problems. To compute a good normalizer for PRM, we propose and study the advantages and disadvantages of a Hit & Run random walk with Dikin ball dilation. We perform this procedure on an ill-conditioned two dimensional simplex to show the Dikin ball Hit & Run random walk mixes much faster than standard Hit & Run random walk. In the last part of this thesis, we conduct computational testing of the PRM on a set of problems from the SDPLIB [3] library derived from control theory and several univariate polynomial problems sum of squares (SOS) problems. Our results reveal that our PRM implementation is effective for problems of smaller dimensions but tends to be ineffective (or even detrimental) for problems of larger dimensions. / by Sai Hei Yeung. / S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Computation for Design and Optimization Program, 2010. / This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Cataloged from student submitted PDF version of thesis. / Includes bibliographical references (p. 67-68). / We study a periodic-reviewed, infinite horizon serial network inventory control problem. The demands in different periods are independent of each other and follow an identical Poisson distribution. Unsatisfied demands are backlogged until they are satisfied by supply units. In each period, there is a per-unit holding cost is incurred for each unit of supply that stays in the system and a per-unit backorder cost is incurred for each unsatisfied unit of demand. The objective of the inventory control policy is to minimize the long-run expected average cost over an infinite horizon. The goal of the thesis is to evaluate the empirical performance of the dual balancing policy and several other variants of cost balancing policies through numerical simulations. The dual-balancing policy is based on two novel ideas: the marginal cost accounting scheme, which assigns to each decision all the costs that are made inevitable after that decision is made; and the cost balancing idea to balance opposing costs. / (cont.) The dual-balancing policy can be modified in several ways to get other cost balancing policies. It has been proven that the dual-balancing policy has a worst-case guarantee of 2 but this does not indicate the empirical performance. An approximately optimal policy is considered as the benchmark to test the quality of the cost balancing policies. In the computational experiments, the dual-balancing policy shows an average error of 7.74% compared to the approximately optimal policy, much better than the theoretical worst-case guarantee. The three variants of cost balancing policies have made significant improvement on the performance of the dual-balancing policy. The accuracy of the dual-balancing policy is also affected by the system parameters. In addition, with high demand rate and long lead times, we have observed several scenarios when the cost balancing policies dominate the approximately optimal policy. / by Qian Yu. / S.M.

Includes bibliographical references (leaf 49). / Thesis (S.M.)--Massachusetts Institute of Technology, Computation for Design and Optimization Program, 2008. / In this thesis, we compare computationally four methods for solving optimization problems under uncertainty: * Robust Optimization (RO) * Adaptive Robust Optimization (ARO) * Data Driven Optimization (DDO) * stochastic Programming (SP) We have implemented several computation experiments to demonstrate the different performance of these methods. We conclude that ARO outperform RO, which has a comparable performance with DDO. SP has a comparable performance with RO when the assumed distribution is the same as the true underlying distribution, but under performs RO when the assumed distribution is different from the true distribution. / by Wang, Yanbo. / S.M.