No description available.

In this research, we designed and implemented optimization search algorithms to facilitate implementation of optimization search software. We provided the design of module interaction graph including modules, ports, and channels. We can map solving algorithms of sub-problems onto behavioral designs incorresponding modules. Finally, they can integrate module¡¦s with channels. Since optimization search algorithms may evolve one to several solutions at the same time, we planned a solution set organization to support designer-planned search strategy. During the optimization process, solutions or sub-solutions should be evaluated and analyzed. Because excessive executive time as commonly spent in replicated evaluation, we planned dynamic programming for reusing evaluation results to reduce replicated evaluation time. Lastly, when evolving new solutions, usually only a small number of decisions are changed. We planed a hierarchical decision representation and maintenance operations to reduce replication of common parts among solutions to further enhance its execution speed.

An optimization strategy for the design of composite shells is investigated. This study differs from previous work in that an advanced analysis package is utilized to provide buckling information on potential designs. The Structural Analysis of General Shells (STAGS) finite element code is used to provide linear buckling calculations for a minimum buckling load constraint. A response surface, spanning the design space, is generated from a set of design points and corresponding buckling load data. This response surface is incorporated into a genetic algorithm for optimization of composite cylinders. Laminate designs are limited to those that are balanced and symmetric. Three load cases and four different variable formulations are examined. In the first approach, designs are limited to those whose normalized in-plane and out-of-plane stiffness parameters would be feasible with laminates consisting of two independent fiber orientation angles. The second approach increases the design space to include those that are bordered by those in the first approach. The third and fourth approaches utilize stacking sequence designs for optimization, with continuous and discrete fiber orientation angle variation, respectively. For each load case and different variable formulation, additional runs are made to account for inaccuracies inherent in the response surface model. This study concluded that this strategy was effective at reducing the computational cost of optimizing the composite cylinders. / Master of Science

A pre-optimizer has been developed which modifies existing turbomachinery blades to create new geometries with improved selected aerodynamic coefficients calculated using a linear panel method. These blade rows can then be further refined using a Navier-Stokes method for evaluation. This pre-optimizer was developed in hopes of reducing the overall CPU time required for optimization when using only Navier-Stokes evaluations. The primary method chosen to effect this optimization is a parallel evolutionary algorithm. Variations of this method have been analyzed and compared for convergence and degree of improvement. Test cases involved both single and multiple row turbomachinery. For each case, both single and multiple criteria fitness evaluations were used.

Algorithms refer to the software programs designed to support problem solving in a wide range of decision domains. Given the Artificial Intelligence (AI) revolution, algorithms have become an integral part of our personal, social, and professional lives. As technology rapidly advances, these algorithms are not only becoming more capable but are also finding a growing array of applications in managerial and consumer decision support. Despite their increasing presence, reactions to algorithms are mixed. While some research highlights a preference for algorithms over human judgment ("algorithm appreciation"), other studies reveal a contrary preference ("algorithm aversion"), where people favor human expertise. This research provides a conceptual framework and empirical evidence regarding factors that may influence preference for algorithmic versus human expert recommendations in business decision contexts. We use experimental psychological methods to investigate how algorithm characteristics, decision-maker psy / Doctor of Philosophy / Amid the AI revolution, algorithms have become central to our personal, social, and professional lives, evolving rapidly in both capability and application. Reactions to these algorithms are mixed: some studies show a preference for algorithms over human judgment, known as "algorithm appreciation," while others reveal a preference for human judgment, or "algorithm aversion." Understanding these preferences is essential. Our research helps to clarify this issue by examining the factors that influence whether people prefer algorithms or human experts in business decisions. Using experimental methods, we explore how algorithm features, decision-maker psychology, and situational factors impact these preferences. We focus on scenarios where algorithms and human experts are presented as competing options rather than complementary ones. Our findings, detailed in two empirical essays, aim to advance marketing literature on algorithms and decision-making, identify future research opportunities, and offer insights for

We consider the scheduling problem of minimizing the average weighted completion time of n jobs with release dates on a single machine. We first study two linear programming relaxations of the problem, one based on a time-indexed formulation, the other on a completiontime formulation. We show their equivalence by proving that a O(n log n) greedy algorithm leads to optimal solutions to both relaxations. The proof relies on the notion of mean busy times of jobs, a concept which enhances our understanding of these LP relaxations. Based on the greedy solution, we describe two simple randomized approximation algorithms, which are guaranteed to deliver feasible schedules with expected objective value within factors of 1.7451 and 1.6853, respectively, of the optimum. They are based on the concept of common and independent a-points, respectively. The analysis implies in particular that the worst-case relative error of the LP relaxations is at most 1.6853, and we provide instances showing that it is at least e/(e - 1) 1.5819. Both algorithms may be derandomized, their deterministic versions running in O(n2 ) time. The randomized algorithms also apply to the on-line setting, in which jobs arrive dynamically over time and one must decide which job to process without knowledge of jobs that will be released afterwards.

We present Glimmer, a new multilevel visualization algorithm for multidimensional scaling designed to exploit modern graphics processing unit (GPU) hard-ware. We also present GPU-SF, a parallel, force-based subsystem used by Glimmer. Glimmer organizes input into a hierarchy of levels and recursively applies GPU-SF to combine and refine the levels. The multilevel nature of the algorithm helps avoid local minima while the GPU parallelism improves speed of computation. We propose a robust termination condition for GPU-SF based on a filtered approximation of the normalized stress function. We demonstrate the benefits of Glimmer in terms of speed, normalized stress, and visual quality against several previous algorithms for a range of synthetic and real benchmark datasets. We show that the performance of Glimmer on GPUs is substantially faster than a CPU implementation of the same algorithm. We also propose a novel texture paging strategy called distance paging for working with precomputed distance matrices too large to fit in texture memory.

No description available.

We present Glimmer, a new multilevel visualization algorithm for multidimensional scaling designed to exploit modern graphics processing unit (GPU) hard-ware. We also present GPU-SF, a parallel, force-based subsystem used by Glimmer. Glimmer organizes input into a hierarchy of levels and recursively applies GPU-SF to combine and refine the levels. The multilevel nature of the algorithm helps avoid local minima while the GPU parallelism improves speed of computation. We propose a robust termination condition for GPU-SF based on a filtered approximation of the normalized stress function. We demonstrate the benefits of Glimmer in terms of speed, normalized stress, and visual quality against several previous algorithms for a range of synthetic and real benchmark datasets. We show that the performance of Glimmer on GPUs is substantially faster than a CPU implementation of the same algorithm. We also propose a novel texture paging strategy called distance paging for working with precomputed distance matrices too large to fit in texture memory.

The expectation-maximization (EM) algorithm is a widely utilized approach to max- imum likelihood estimation in the presence of missing data, this thesis focuses on its application within the model-based clustering framework. The performance of the EM algorithm can be highly dependent on how the algorithm is initialized. Several ways of initializing the EM algorithm have been proposed, however, the best method to use for initialization remains a somewhat controversial topic. From an attempt to obtain a superior method of initializing the EM algorithm, comes the concept of using multiple existing methods together in what will be called a `voting' procedure. This procedure will use several common initialization methods to cluster the data, then a nal starting ^zig matrix will be obtained in two ways. The hard `voting' method follows a majority rule, whereas the soft `voting' method takes an average of the multiple group memberships. The nal ^zig matrix obtained from both methods will dictate the starting values of ^ g; ^ g; and ^ g used to initialize the EM algorithm.