Resource management for a data center with multiple job classes is investigated in this thesis. We focus on strategies for allocating resources to an application mix such that the service level agreements (SLAs) of individual applications are met. A performance model with two interactive job classes is used to determine the smallest number of processor nodes required to meet the SLAs of both classes. For each class, the SLA is specified by the relationship: Prob(response time≤x)≥y. Two allocation strategies are considered: shared allocation (SA) and dedicated allocation (DA). For the case of FCFS scheduling, analytic results for response time distribution are used to develop a heuristic algorithm that determines the allocation strategy (SA or DA) that requires fewer processor nodes. The effectiveness of this algorithm is evaluated over a range of operating conditions. The performance of SA with non-FCFS scheduling is also investigated. Among the scheduling disciplines considered, a new discipline called probability dependent priority (PDP) is found to have the best performance in terms of requiring the smallest number of nodes. Furthermore, we extend our heuristic algorithm for FCFS to three job classes. The effectiveness of this extended algorithm is evaluated. As to priority scheduling, the performance advantage of PDP is also confirmed for the case of three job classes.

Resource management for a data center with multiple job classes is investigated in this thesis. We focus on strategies for allocating resources to an application mix such that the service level agreements (SLAs) of individual applications are met. A performance model with two interactive job classes is used to determine the smallest number of processor nodes required to meet the SLAs of both classes. For each class, the SLA is specified by the relationship: Prob(response time≤x)≥y. Two allocation strategies are considered: shared allocation (SA) and dedicated allocation (DA). For the case of FCFS scheduling, analytic results for response time distribution are used to develop a heuristic algorithm that determines the allocation strategy (SA or DA) that requires fewer processor nodes. The effectiveness of this algorithm is evaluated over a range of operating conditions. The performance of SA with non-FCFS scheduling is also investigated. Among the scheduling disciplines considered, a new discipline called probability dependent priority (PDP) is found to have the best performance in terms of requiring the smallest number of nodes. Furthermore, we extend our heuristic algorithm for FCFS to three job classes. The effectiveness of this extended algorithm is evaluated. As to priority scheduling, the performance advantage of PDP is also confirmed for the case of three job classes.

As health care costs continue to increase at rates higher than the general inflation rate, there is increased focus on controlling health care expenditures in the public and private sectors. In particular, there is a compelling need for more creative and informed allocation decisions for limited government public health funds. This thesis suggests several methods for better forecasting the demand for health care and allocating health care resources more efficiently. First, productivity of dental sealant programs is studied and suggestions are made for increased efficiency. Using simulation and data from several states programs, guidelines are offered for optimal programs based on program size, distance to site, and practice act requirements. We find that under most conditions, it is better to carry an extra dental assistant to every program. The cost of satisfying practice act requirements is also quantified. Second, a model for allocating health resources to Community Health Centers (CHCs) is provided. Using the state of Georgia as a prototype, local estimation is used to forecast county insurance types, disease prevalence, and likelihood of using a clinic. Then, the optimal locations and service portfolios to be offered under financial constraints are determined using a developed mixed-integer programming model. Finally, shortcomings in current Markovian modeling of chronic disease are analyzed. Common forecasting techniques can overestimate or underestimate the population in need of care, as illustrated by analytic results and an example with lung cancer data. The chapter presents suggestions for improving such modeling. Each of these issues affect the planning models for scarce resources in health care, and improving those models can positively impact utilization of those services. Through this research, models are presented that can positively impact public health decisions in coming years, particularly those for growing high-risk and low-income groups.

The increased prevalence of childhood obesity is a major concern for society. This study aims at exploring the influence of the parents (especially parental time allocation choices) on children’s obesity-related health outcomes and examining the potential differences between the fathers’ and the mothers’ marginal effects. A household with two parents and one child is modeled. The household production theory and the collective household modeling structure are combined. The model treats the mother, the father and the child as three separate agents with individual preferences. The two parents’ interaction is modeled within the collective model framework by assuming that they will reach Pareto efficient resource allocation between them. In order to capture the dynamics between parents and the child, parents-child interaction is modeled in a two-stage Stackleberg game structure where the child is allowed to have certain decision choices of his/her own. This game structure allows us to explore the parental influence on the child’s health outcomes while allowing the child to have influencing power in the household decision-making process. Based on this theoretical model, a general triangular system with one child’s health production equation and five health inputs demand equations is derived and estimated. The empirical estimation is performed for three systems: pooled model, the younger children model (of age 9 to 11), and the older children model (of age 13 to 15). The empirical results show mother-related variables show more influence on the child’s Body Mass Index (BMI) outcomes compared to father-related variables: mothers’ BMI and mothers’ work-to-home stress spillover are positively related to their children’s BMI while mothers’ time spent with their children is negatively related to their children’s BMI. There exists a complementary relationship between mothers’ income and fathers’ food preparation time. In the older children model, mothers’ own income increases tend to decrease their time spent with their children. The main contribution of this study is that it develops a general theoretical framework to capture the dynamics in parents-child interaction. Based on this theoretical model, empirical analysis and future work can be conducted in a theoretically consistent way.

Thesis (Ph. D.)--University of Chicago, Dept. of Economics, June 2001. / Includes bibliographical references. Also available on the Internet.

Dissertation (Ph.D.)--Auburn University, / Abstract. Includes bibliographic references (p.141-146).

Thesis (M.B.A.)--Naval Postgraduate School, December 2003. / "MBA professional report"--Cover. Thesis advisor(s): Joseph G. San Miguel, Don E. Summers. Includes bibliographical references (p. 35). Also available online.

published_or_final_version / Economics and Finance / Doctoral / Doctor of Philosophy

A desire to advance humanity's presence in space prompts the need for improved technology to send crew to places such as the Moon, Mars, and nearby asteroids. Safely placing a crewed vehicle on and in any landing condition requires a design decision regarding the distribution of responsibilities between the crew and automation. In this thesis, a cognitive process model is used to determine the necessary automated functionality to support astronaut decision making. Current literature lacks sufficient detailed knowledge regarding astronaut decision making during this task and observations of astronauts landing on the Moon are not readily available. Therefore, a series of human-in-the-loop experiments, one of which was conducted with the NASA Astronaut Office at Johnson Space Center, have been conducted to examine the changes in performance due to differing function allocations, trajectory profiles, and scenario operations. The data collected in the human-in-the-loop study has provided empirical data that has informed the cognitive process model, the requirements analysis, and provided insight regarding cockpit display usage and information needs. The proposed system requirements include design guidance for assisting astronauts during both nominal and off-nominal landing scenarios.

No description available.