Over the years analysts have used the EM algorithm to obtain maximum likelihood estimates from incomplete data for various models. The general algorithm admits several appealing properties such as strong global convergence; however, the rate of convergence is linear which in some cases may be unacceptably slow. This work is primarily concerned with applying Anderson acceleration to the EM algorithm for Gaussian mixture models (GMM) in hopes of alleviating slow convergence. As preamble we provide a review of maximum likelihood estimation and derive the EM algorithm in detail. The iterates that correspond to the GMM are then formulated and examples are provided. These examples show how faster convergence is experienced when the data are well separated, whereas much slower convergence is seen whenever the sample is poorly separated. The Anderson acceleration method is then presented, and its connection to the EM algorithm is discussed. The work is then concluded by applying Anderson acceleration to the EM algorithm which results in reducing the number of iterations required to obtain convergence.

The multicommodity flow problem involves shipping multiple commodities simultaneously through a network so that the total flow over each edge does not exceed the capacity of that edge. The concurrent flow problem also associates with each commodity a demand, and involves finding the maximum fraction z, such that z of each commodity's demand can be feasibly shipped through the network. This problem has applications in message routing, transportation, and scheduling problems. It can be formulated as a linear programming problem, and the best known solutions take advantage of decomposition techniques for linear programming. Often, quickly finding an approximate solution is more important than finding an optimal solution. A solution is epsilon-optimal if it lies within a factor of (1+epsilon) of the optimal solution. We present a combinatorial approximation algorithm for the concurrent flow problem. This algorithm consists of finding an initial flow, and gradually rerouting this flow from more to less congested paths, until an epsilon-optimal flow is achieved. This algorithm theoretically runs much faster than linear programming based algorithms.

Estendemos o trabalho de Vincent Rijmen e Joan Daemen na estratégia de trilha larga, uma metodologia de projeto para primitivas criptográficas simétricas eficientes e demonstravelmente resistentes às técnicas de criptanálise diferencial e linear. Preocupamo-nos principalmente com a melhoria na eficiência de primitivas projetadas de acordo com a estratégia de trilha larga. Investigamos duas linhas distintas de pesquisa: a aplicabilidade da técnica de bitslicing à implementação em software de primitivas baseadas na estratégia de trilha larga; e o projeto de S-boxes estruturadas com implementação eficiente em hardware e bitslicing, e especificamente, o uso de S-boxes invariantes por rotação, que exibem propriedades vantajosas para implementação. Também implementamos e otimizamos algumas primitivas criptográficas em plataformas de software selecionadas, para substanciar e aprimorar as afirmações de eficiência da estratégia de trilha larga. Ademais, aplicamos nosso conhecimento e técnicas propostas ao projeto de novas primitivas criptográficas altamente eficientes, em particular a função de hash MAELSTROM-0 e a cifra de bloco legada FUTURE. / We extend the work of Vincent Rijmen and Joan Daemen on the Wide Trail strategy, a design methodology for symmetric-key cryptographic primitives which are efficient and provably secure against differential and linear cryptanalysis. We concern ourselves mainly with improving the efficiency of primitives designed according to the Wide Trail strategy. To that end, we investigate two distinct lines of research: the applicability of the bitslicing technique to the software implementation of primitives based on the Wide Trail strategy; and the design of structured S-boxes with efficient implementation in hardware and bitslicing, and specifically, the use of rotation-symmetric S-boxes, which exhibit advantageous implementation properties. We also perform general implementation and optimization work on selected software platforms, to further realize the claims of efficiency of the Wide Trail strategy. Additionally, we apply our expertise and proposed techniques to the design of new highly-efficient cryptographic primitives, in particular the hash function MAELSTROM-0 and the legacy-level block cipher FUTURE.

"Concepts in manufacturing such as CIMS (Computer Integrated Manufacturing Systems), JIT (Just In Time), Lean Production, Virtual Manufacturing, and Flexible Fixturing have been proposed to meet the fundamental requirements of manufacturing - decrease the cost and satisfy the needs of customers. Fast fixture generation and fixture reusability are essential in the current manufacturing environment. The dissertation focuses on the models, methods, and algorithms for fixture synthesis and variation that satisfy the functional requirements specified by on-site industrial engineers. With the reusability of a fixture base combined with variation of other fixture components, fixture configuration can be rapidly adapted and accommodated to the new workpiece. The dissertation presents methods and algorithms for fixture base synthesis, which directly result in fixture reusability. Optimization functions are derived based on engineering requirements due to the mass production nature of automotive parts. Specific optimization algorithms are developed and their complexities, compared to other alternatives, are comprehensively evaluated according to different optimization functions. The fixture variation and reusability provide an engineering tool to rapidly generate and validate fixtures in production planning stage. It applies scientific reasoning methodology in combination with best knowledge of fixture designs, which heavily relies on designers' manufacturing knowledge and experience. It also provides means to bridge the gap between CAD and CAM integration and therefore reduces the new product and production development cycle time and cost while maintaining the quality of fixtures."

We address the issue of shuffling loads in Automated Storage/Retrieval Systems (AS/RS) in this paper. The objective is to pre-sort the loads into any specified locations in order to minimize the response time of retrievals. 1D, 2D and 3D AS/RS racks have been designed in order to achieve the shuffling efficiently. The shuffling algorithms are described in detail. The response time of retrieval, the lower and upper bounds of energy consumption are also derived. Results of the analysis and numerical experiments show that the shuffling algorithms are quite efficient. / Singapore-MIT Alliance (SMA)

A simple but effective evolutionary algorithm is proposed in this paper for solving complicated optimization problems. The new algorithm presents two hybridization operations incorporated with the conventional genetic algorithm. It takes only 4.1% ~ 4.7% number of function evaluations required by the conventional genetic algorithm to obtain global optima for the benchmark functions tested. Application example is also provided to demonstrate its effectiveness. / Singapore-MIT Alliance (SMA)

Researchers have applied genetic algorithms (GAs) and other heuristic optimization methods to perform truss optimization in recent years. Although a substantial amount of research has been performed on the optimization of truss member sizes, nodal coordinates, and member connections, research that seeks to simultaneously optimize the topology, geometry, and member sizes of trusses is still uncommon. In addition, most of the previous research is focused on the problem domains that are limited to a structured domain, which is defined by a fixed number of nodes, members, load locations, and load magnitudes. The objective of this research is to develop a computational method that can design efficient roof truss systems. This method provides an engineer with a set of near-optimal trusses for a specific unstructured problem domain. The unstructured domain only prescribes the magnitude of loading and the support locations. No other structural information concerning the number or locations of nodes and the connectivity of members is defined. An implicit redundant representation (IRR) GA (Raich 1999) is used in this research to evolve a diverse set of near-optimal truss designs within the specified domain that have varying topology, geometry, and sizes. IRR GA allows a Pareto-optimal set to be identified within a single trial. These truss designs reflect the tradeoffs that occur between the multiple objectives optimized. Finally, the obtained Pareto-optimal curve will be used to provide design engineers with a range of highly fit conceptual designs from which they can select their final design. The quality of the designs obtained by the proposed multi-objective IRR GA method will be evaluated by comparing the trusses evolved with trusses that were optimized using local perturbation methods and by trusses designed by engineers using a trial and error approach. The results presented show that the method developed is very effective in simultaneously optimizing the topology, geometry, and size of trusses for multiple objectives.

This thesis proposes a clustering algorithm for vehicular communication networks. A novel clustering metric and an improved clustering framework are introduced. The novel clustering metric, network criticality, is a global metric on undirected graphs which quantifies the robustness of the graph against changes in environmental parameters, and point-to-point network criticality is also defined to measure the resistance between different points of a graph. We localize the notion of network criticality for a node of a vehicular network which can potentially be promoted as the cluster header. We use the localized notion of node criticality in conjunction with a universal link metric, Link Expiration Time (LET), to derive a clustering algorithm for the vehicular network. We employ a distributed multi-hop clustering algorithm based on the notion of network criticality. Simulation results show that the proposed clustering algorithm forms a more robust cluster structure.

This thesis proposes a clustering algorithm for vehicular communication networks. A novel clustering metric and an improved clustering framework are introduced. The novel clustering metric, network criticality, is a global metric on undirected graphs which quantifies the robustness of the graph against changes in environmental parameters, and point-to-point network criticality is also defined to measure the resistance between different points of a graph. We localize the notion of network criticality for a node of a vehicular network which can potentially be promoted as the cluster header. We use the localized notion of node criticality in conjunction with a universal link metric, Link Expiration Time (LET), to derive a clustering algorithm for the vehicular network. We employ a distributed multi-hop clustering algorithm based on the notion of network criticality. Simulation results show that the proposed clustering algorithm forms a more robust cluster structure.

Hard combinatorial optimization (CO) problems pose challenges to traditional algorithmic solutions. The search space usually contains a large number of local optimal points and the computational cost to reach a global optimum may be too high for practical use. In this work, we conduct a comparative study of several state-of-the-art metaheuristic algorithms for hard CO problems solving. Our study is motivated by an industrial application called the Fertilizer Blends Optimization. We focus our study on a number of local search metaheuristics and analyze their performance in terms of both runtime efficiency and solution quality. We show that local search granularity (move step size) and the downhill move probability are two major factors that affect algorithm performance, and we demonstrate how experimental tuning work can be applied to obtain good performance of the algorithms. <p>Our empirical result suggests that the well-known Simulated Annealing (SA) algorithm showed the best performance on the fertilizer problem. The simple Iterated Improvement Algorithm (IIA) also performed surprisingly well by combining strict uphill move and random neighborhood selection. A novel approach, called Delivery Network Model (DNM) algorithm, was also shown to be competitive, but it has the disadvantage of being very sensitive to local search granularity. The constructive local search method (GRASP), which combines heuristic space sampling and local search, outperformed IIA without a construction phase; however, the improvement in performance is limited and generally speaking, local search performance is not sensitive to initial search positions in our studied fertilizer problem.