Stochastic Petri nets (SPNs) and extensions are a popular method for evaluating a wide variety of systems. In most cases, the interesting measures regarding the system's characteristics can be defined at the net level by means of reward variables. Depending on the measures, these net-level reward models are solved either by first generating a state-level reward model or by directly generating paths from the net-level description. In this thesis, we propose algorithms for the generation of state-level reward models as well as for directly obtaining solutions from net-level reward models when the net-level reward models are specified as stochastic activity networks (SANs) with "step-based reward structure." Moreover, we propose algorithms for computing the expected value and the probability distribution function of a reward variable at specified time instants, and for computing the probability distribution function of reward accumulated during a finite interval. The interval may correspond to the mission period in a mission-critical system, the time between scheduled maintenances, or a warranty period; whereas the time instants may be critical instances during these intervals. The proposed algorithms avoid the construction of state-level representations and the memory growth problems experienced when applying previous approaches to large models. Furthermore, we study the effect of workload on the availability and response time of voting algorithms. Voting algorithms are a popular way to provide data consistency in replicated data systems. Many models have been made to study the degree to which replication increases the availability of data, and some have been made to study the cost incurred in maintaining consistency. However, little work has been done to evaluate the time it takes to serve request, accounting for server and network failures, or to determine the effect of workload on these measures. In this thesis, we use stochastic activity networks (SANs) to study the effect of work load on availability and mean response time of two variant models of a replicated file system to maintain data consistency, one using a static voting algorithm, the other using a dynamic voting algorithm.

The Markovian Arrival Process (MAP) is a tractable, versatile class of Markov renewal processes which has been extensively used to model arrival (or service) processes in queues. This thesis mainly deals with the first two moment matrices of the counts for the MAP. We derive asymptotic expansions for these two moment matrices and also derive efficient and stable algorithms to compute these matrices numerically. Simpler expressions for some of the classical mathematical descriptors of the superposition of independent MAPs also are derived.

This thesis proposes a methodology for factory layout synthesis that combines the symbolic and numerical approaches. The layout model is represented as a loosely packed arrangement of rectangles with sides parallel to two orthogonal directions. Solutions must satisfy a set of qualitative and quantitative constraints and must optimize material handling costs. Qualitative constraints are specified using a rule based approach. Quantitative constraints and material handling costs are incorporated into a numerical, non linear-optimization problem. Optimal solutions are generated by inserting one device at the time in all admissible locations on the floorplan. The resulting partial layouts are optimized with respect to material handling costs. Then, the best partial solution obtained so far is the starting point for the next object insertion. The process stops when all devices have been placed. To illustrate the methodology, three case studies are presented.

Throughout much of the past century, manufacturing efficiencies were gained by constructing systems from independently designed and optimized tasks. Recent theories and practice have extolled the virtues of team-based practices that rely on human flexibility and empowerment to improve integrated system performance. The formation of teams requires consideration of innate tendencies and interpersonal skills as well as technical skills. In this project we develop and test mathematical models for formation of effective human teams. Team membership is selected to ensure sufficient breadth and depth of technical skills. In addition, measures of worker conative tendencies are used along with empirical results on desirable team mix to form maximally effective teams.

<p> Low back pain (LBP) is the most prevalent work-related musculoskeletal disorder. Occupational risk factors have been studied for current ergonomic prevention strategies; however, other underlying mechanisms may exist since not all workers performing the same task develop the same severity. Previous research has identified personal and psychosocial risk factors that also contribute to LBP. Research quantifying the interactive effects of the various personal, psychosocial and occupational factors is limited, along with research on the effect of risk factor combinations on LBP severity. </p><p> The objectives of this study were to: 1) study the various factors that are known to be involved in low back pain and analyze interactions, and 2) develop a model to predict low back pain and validate it. In order to address these objectives, 2 studies were conducted. </p><p> The first study investigated the effects of various personal, genetic, occupational and psychosocial factors on two subjective LBP severity ratings: Oswestry Disability Index (ODI) and a Visual Analog Scale (VAS), and three physician-based ratings: MRI severity, canal stenosis and nerve impingement. Personal and psychosocial factors, in addition to occupational factors, were found to significantly affect the severity ratings. </p><p> The second study involved building predictive models of LBP severity for each risk factor category as well as a combined risk factor model. Results showed that the combined risk factor models considering interaction effects both within and across risk factor categories were significantly better in predicting severity ratings than the individual models. However, validation conducted using 5 random samples showed inconsistent accuracies. Results obtained may help to develop a more reliable way to predict and, hence, prevent chronic LBP.</p>

<p> The purpose of this quantitative research study presents the correlations between education, years of experience, and Six Sigma experience versus an individual&rsquo;s compensation because of training and participating in the Six Sigma Methodology with regards to organizational management theory and leadership. The literature review revealed the historic foundations of Total Quality Management, Six Sigma and Lean aspects of participating in Six Sigma. The participant population consisted of Green Belts, Black Belts, Master Black Belts and Champions from several different industries within the United States and abroad. The participant data were collected through a comprehensive questionnaire with the following categories: (a) demographics; (b) Six Sigma training; (c) industry experience; (d) Six Sigma project completion; (e) compensation; (f) Lean Six Sigma tools frequency of use. The Spearman&rsquo;s rho results of the study indicated that Six Sigma training and participation is statistically significant and an important benefit to participants as compared to years of experience and education. An individual trained in the Six Sigma Methodology earns more to significantly more over the course of their career after participating and practicing Six Sigma. The study provides leadership with an understanding of the expected Six Sigma participant salaries and bonus percentage after training and successfully practicing the Six Sigma Methodology, as well as a brief discourse and recommendations on Six Sigma implementation within matrix organizations through a transformational leadership style.</p>

Web designers attempt to draw attention to important links by making them distinctive. However, when users are asked to find specific items, they often overlook these distinctive banners. The irony of this phenomenon I call "banner blindness" is that the user who really wants to find the information the designer has highlighted is not likely to do so. In the experiments reported here, banner blindness was investigated under controlled conditions. Banners located higher on the page and therefore farther from other links were missed more often than banners located lower on the page and closer to the other links. Banners were missed more often when located on pages containing links to categories than when located on pages with links to specific items. Users rarely noticed banners when clicking the banner was not required to accomplish a task. Banner blindness occurred with several types of distinctive links---graphical banners that resembled advertisements, large plain-text banners and small plain-text banners that were very unlike advertisements. Increasing the perceptual grouping between the banner and the "menu" of hyperlinks helped users notice the banners only slightly more often. Adding animation to graphical banners did not help mitigate the effect. Users searching for specific information seem to focus exclusively on the link-rich areas of the page and do not notice distinctive items outside of that area. The last two experiments in this research focused on emphasizing one item within a menu of search-engine "hits." Three types of emphasis were used. Very large text caused a slight banner-blindness effect. Subtly large text had no effect at all. Highlighting one menu item by giving it a brightly-colored background did not cause banner blindness. In fact, it attracted the attention of users: users were more likely to select the highlighted item and did so more quickly. This type of color highlighting was most effective when it emphasized the first item in the menu. It was slightly less effective when it emphasized items in the middle of the menu.

Electromagnetic levitation, which can provide rapid heating and melting, homogeneity of melt and minimal specimen contamination, is an important branch of containerless processing. The longitudinal electromagnetic levitator is a new type of levitation device, which was invented recently and has a potential to become a containerless manufacturing processing tool. It has some unique advantages, such as good visual access to the sample, capability to support multiple samples, large loads and cylindrical shape sample availability. In this thesis, a brief review of the history and application of electromagnetic levitation is presented. Then the detailed theoretical analysis coupled with experimental work validating the theoretical models of the longitudinal electromagnetic levitator are presented. First, a new electric current model is introduced, which is more appropriate for the computation of the electromagnetic force field in the levitated specimen. Based on this new model, the essential equations for the electromagnetic field and the lifting force field for a cylindrical sample are derived, the current density distribution and the averaged power in the sample are analyzed. Additionally, both lifting force and lifting capacity for the longitudinal levitator are investigated analytically, and compared with experimental data with good agreement. These theoretical predictions can be used to design longitudinal levitators, to select suitable material for levitation, and to provide the framework for further investigation of materials processing using the longitudinal levitator. In addition, temperature distribution simulation for the sample levitated in the longitudinal electromagnetic levitator is implemented by analytical and numerical ways. Isothermal case, steady state and lumped system are discussed respectively as some special cases. The exact solution and numerical simulation of the temperature distribution for the levitated sample are compared with good agreement. The flow motion within the levitated sample and the numerical simulation of the temperature distribution with flow convection has also been investigated. The results provide important information of the levitation phenomena which are very useful for scientific and engineering applications, especially for materials processing.

Injection molding is a cyclic process used for the fabrication of thermosetting and thermoplastic articles. The thermoplastic polymer is melted and injected into the cavity, where it is molded under pressure and ejected after solidification. The amount of polymer mass contained in the cavity is the part weight. The control of part weight is important to ensure quality injection molded parts. The part weight is determined by the state of the polymer at the time the cavity gate freezes. The bulk temperature and the peak cavity pressure at the gate are used to characterize this state. / Measuring internal polymer temperature profiles in the injection mold cavity during molding is extremely difficult. This work presents a method which combines measurements of cavity surface temperatures, cavity pressure, and on-line calculations for estimating temperature profiles inside the cavity. These profiles are then used to estimate the bulk polymer temperature. Fitting the cycle-to-cycle values of bulk polymer temperature and peak pressure to a Tait equation of state yields a model to predict part weights. / The part weight is controlled through the use of a control strategy for the cavity pressure and the part weight model, together with the on-line estimation of the bulk temperature. A self-tuning algorithm with an observer is employed for controlling the cavity pressure time profile to a set point trajectory. The dynamics and control of the bulk temperature are also studied.

The capacitated multi-item lot sizing problem is a model which aims at scheduling production of several products over a finite number of periods, while minimizing production costs, holding inventory costs and setup costs subject to demand and capacity constraints. These costs may vary for each product and each period and are all linear. Our model includes setup times for each product. / We compare two approaches: a classic Lagrangean relaxation of the capacity constraints and a Lagrangean decomposition by variable splitting. In both cases, the Lagrangean multipliers are updated with an interior point cutting plane technique. The results show: (1) The superiority of the interior point method over the commonly used subgradient optimization in terms of accuracy at termination, number of iterations and ease of utilization. (2) The better quality of the bounds obtained by the Lagrangean decomposition by variable splitting over the Lagrangean relaxation.