Computer simulations are routines programmed to imitate detailed system operations. They are utilized to evaluate system performance and/or predict future behaviors under certain settings. In complex cases where system operations cannot be formulated explicitly by analytical models, simulations become the dominant mode of analysis as they can model systems without relying on unrealistic or limiting assumptions and represent actual systems more faithfully. Two main streams exist in current simulation research and practice: discrete event simulation and agent-based simulation. This dissertation facilitates the marriage of the two. By integrating the agent-based modeling concepts into the discrete event simulation framework, we can take advantage of and eliminate the disadvantages of both methods.<br><br>Although simulation can represent complex systems realistically, it is a descriptive tool without the capability of making decisions. However, it can be complemented by incorporating optimization routines. The most challenging problem is that large-scale simulation models normally take a considerable amount of computer time to execute so that the number of solution evaluations needed by most optimization algorithms is not feasible within a reasonable time frame. This research develops a highly efficient evolutionary simulation-based decision making procedure which can be applied in real-time management situations. It basically divides the entire process time horizon into a series of small time intervals and operates simulation optimization algorithms for those small intervals separately and iteratively. This method improves computational tractability by decomposing long simulation runs; it also enhances system dynamics by incorporating changing information/data as the event unfolds. With respect to simulation optimization, this procedure solves efficient analytical models which can approximate the simulation and guide the search procedure to approach near optimality quickly.<br><br>The methods of agent-based discrete event simulation modeling and evolutionary simulation-based decision making developed in this dissertation are implemented to solve a set of disaster response planning problems. This research also investigates a unique approach to validating low-probability, high-impact simulation systems based on a concrete example problem. The experimental results demonstrate the feasibility and effectiveness of our model compared to other existing systems.

Three interrelated aspects of the product lifecycle decision making are studied from the perspective of sustainability: waste reduction in manufacturing, green product deployment strategies and product upgrade investment decision making. Mathematical models are developed to assist stakeholders in making the rational decisions at corresponding product lifecycle stages.

Several ethical decision making models have been developed over the last twenty years. Past research has attempted to evaluate these models by assessing numerous factors potentially linked to the decision process involving ethical issues. Past research studying ethical decision making in organizations has focused on the business perspective and on individual decision making. Little empirical research has focused on teams ethical decision making in engineering and none (to the authors knowledge) have studied the process of ethical decision making by engineers. For this research two primary models have been adopted: Joness Synthesis of Ethical Decision Making model and the Harris, Pritchard, and Rabins (HPR) Model widely used in engineering. These models were combined along with factors cited in the literature to form a proposed Ethical Decision Making in Engineering Model. Using this model an experimental study involving both individuals and teams of engineering students solving two ethical dilemmas of different moral intensity was used to: (1) investigate whether engineering student teams make better decisions than individual engineering students, (2) evaluate the processes used by the individuals and teams to resolve the dilemmas, (3) and assess variables that potentially affect the quality of the resolution and the quality of the decision process. From this research, the analysis of the team decision making process and its outcomes has enabled the researcher to identify key factors that play a role in engineering ethical decision making, as well as identify potential improvement areas for engineering ethics education. In general, students who have had an engineering ethics course perform better (in teams or as individuals) than students who did not have engineering ethics course for an engineering dilemma with moderate moral intensity; and teams outperformed individuals on the Resolution attribute and spent more time on Analysis and Recognition of Dilemma attributes. Further, the derived regression analysis models showed that having had an engineering ethics course, working in teams, work experience, being female, the type of engineering major, and the dilemmas moral intensity are significant predictors of the overall Resolution as measured by the report quality.

In todays rapidly changing global world, the sustainability of an organization depends not only upon its financial performance, but also upon its environmental and social performance. It is suggested that policy makers, and corporate and engineering managers integrate economic, environmental and social objectives i.e., the triple bottom line (TBL) into their overall strategic plan and consider these objectives in their decision making. Investment planning and capital budgeting decisions play a critical role in aligning an organization with its economic, environmental and social strategic objectives. This research introduces a new decision making tool that integrates both financial and non-financial performance measures into the process of investment planning and capital budgeting via the TBL. It makes use of stakeholder theory for group decision making, analytic network process (ANP) as a decision support tool and stochastic linear programming to create an optimal investment portfolio. This new tool evaluates and prioritizes a set of projects and creates a long-term balanced investment portfolio based upon the perspectives and priorities of the stakeholder groups and decision makers. It can assist decision makers with developing and making proactive decisions which support the strategy of their organization with respect to economic, environmental and social issues, ensuring the sustainability of their organization in the future. To create a sustainability culture both in academia and business environment, and to encourage communities for sustainable development, a real life application of the developed tool is provided through coordination with Sustainable Pittsburgh and Cranberry Township business leaders.

In 2009, almost 1,500 Americans died of end-stage liver disease (ESLD), which is the twelfth leading cause of death in the U.S. As liver transplantation is the only possible therapy for ESLD and there is a considerable difference between the number of donated organs and patients, it is important to manage donor-patient match and investigate alternative treatments to transplantation. Every patient lists in at least one waiting list (OPO) in order to be eligible for a donated organ. However, patients may list in additional OPOs. This practice is called multiple listing. Currently, multiple listing is one of the most debated topics in organ allocation. Although transplantation is a successful procedure, it may not be available on time due to the massive shortage of donated organs. Therefore, an alternative therapy to transplantation is needed. Liver Assist Devices (LADs) are an emerging therapy for ESLD that aim to stabilize a patient until transplantation or her own organ recovers. In this dissertation, we discuss three models that are related to ESLD. In the first model, we optimize the three-stage decision process faced by a single patient. The patient decides her geographic location, in which OPOs to multiple list, and which organ offers to accept. This problem is formulated as a continuous-time Markov Decision Process (MDP). We derive structural properties of this model and solve it using clinical data. The second model analyzes multiple listing from the societal perspective. Utilizing an existing simulation of the U.S. liver allocation system, we give every patient the flexibility to multiple list. Therefore, we evaluate the effects of multiple listing on every wait-listed patient, rather than on a single patient. We also study the same problem where multiple listing is a more widespread practice in the U.S. The third model considers a hypothetical system in which an internal LAD is available. So, in addition to the liver accept/reject decision, patients can decide to accept an LAD. This model aims to help manufacturers by estimating potential demand for an LAD. We model this problem as a discrete-time MDP and give sufficient conditions under which an LAD will be worthwhile.

Based on an overall consideration of the principles and characteristics in designing a retail area layout, this research is the first work to integrate aisle structure design, block layout, which is specific department placement, and intra-block departmental layout, which is detailed fixture layout design, as a whole process. The main difference between previous research and this proposed research is the formulation of mathematical models that can be specified applied in the retail sector. Unlike manufacturing, in retail environments, the design objective is profit maximization. This is accomplished by maximizing the area exposure, optimizing the adjacency preference of all departments, and adjusting the intra-block layout and evaluating the effectiveness of layout design.

A focus of engineering education is to prepare future engineers with problem solving, design and modeling skills. In engineering education, the former two skill areas have received copious attention making their way into the ABET criteria. Modeling, a representation containing the essential structure of an event in the real world, is a fundamental function of engineering, and an important academic skill that students develop during their undergraduate education. Yet modeling process remains under-investigated, particularly in engineering, even though there is an increasing emphasis on modeling in engineering schools (Frey 2003). Research on modeling requires multiple perspectives, that of cognition, affect, and knowledge expansion. In this dissertation, the relationship between engineering modeling skill and students' cognitive backgrounds including self-efficacy, epistemic beliefs and metacognition is investigated using model-eliciting activities (MEAs). The impact of each cognitive construct on change in modeling skills was measured using a growth curve model at the sophomore level, and ordinary least squares regression at the senior level. Findings suggest that self-efficacy, through its direct and indirect (moderation or interaction term with time) impact, influences the growth of modeling abilities of an engineering student. When sophomore and senior modeling abilities are compared, the difference can be explained by varying self-efficacy levels. Epistemology influences modeling skill development such that the more sophisticated the student beliefs are, the higher the level of modeling ability students can attain, after controlling for the effects of conceptual learning, gender and GPA. This suggests that development of modeling ability may be constrained by the naivet&#xE9; of one's personal epistemology. Finally, metacognition, or 'thinking about thinking', has an impact on the development of modeling strategies of students, when the impacts of four metacognitive dimensions are considered: awareness, planning, cognitive strategy and self-checking. Students who are better at self-checking show higher growth in their modeling abilities over the course of a year, compared to students who are less proficient at self-checking. The growth is moderated by the cognitive strategy and planning skills of the student. Therefore, inherent metacognitive abilities of students can positively affect the growth of modeling ability.

This thesis consists of three essays concerning the use of optimization techniques to solve four problems in the fields of data mining, computational biology, and sustainable energy devices. To the best of our knowledge, the particular problems we discuss have not been previously addressed using optimization, which is a specific contribution of this dissertation. In particular, we analyze each of the problems to capture their underlying essence, subsequently demonstrating that each problem can be modeled as a nonlinear (mixed) integer program. We then discuss the design and implementation of solution techniques to locate optimal solutions to the aforementioned problems. Running throughout this dissertation is the theme of using mixed-integer programming techniques in conjunction with context-dependent algorithms to identify optimal and previously undiscovered underlying structure.

Industry leaders in todays global market strive for continuous improvement in order to remain competitive. One method used by firms for cutting costs and improving efficiency is Design for Supply Chain (DFSC). The objective of this methodology is to design the supply chain in parallel to designing or redesigning a new product. Risk is an inherent element of this DFSC process. Although supply chain risk models and new product development risk models are available, there are few models that consider the combined effect of risk to product development and the supply chain. A gap in the body of knowledge could be filled by a DFSC and risk model that looks at design, supply chain and risk concurrently. This research develops such a model and tests it on two data sets. The most critical risks to incorporate in the model were found through a review of the literature and a survey of industry experts. The model consists of two components. The first component is a Mixed Integer Programming (MIP) model which makes the DFSC decisions while simultaneously considering time-to-market risk, supplier reliability risk and strategic exposure risk. The results from the MIP are then used in the second model component which is a discrete event simulation. The simulation tests the robustness of the MIP solution for supplier capacity risk and demand risk. When a decision maker is potentially facing either of these risks the simulation shows whether it is best to use an alternative solution or proceed with the MIP solution. The model provides analytical results to be used by decision makers, but also allows decision makers to use their own judgment to select the best option for overall profitability. It is shown that the DFSC model with risk is a powerful decision making tool.

A critical factor in increasing the widespread adoption of Radio Frequency Identification (RFID) technology for different supply chain applications is the ability to achieve a high level of read accuracy. The read accuracy is dependent on the size of the region that receives sufficient power from the reader. While most current research considers the powering region of a reader to be determined only by its read range, in reality read accuracy can be complicated by such issues as polarizations and the relative orientations of reader antennas and tags. In particular, when tag positions are not fixed, the specific placement of reader antennas and their interaction with the polarization and the orientation of the tags can have a significant effect on the success of the interrogation processes. This research uses Friis equation for both the forward link and the backward link to explicitly consider orientations and polarizations while addressing the problem of optimizing the locations of a set of reader antennas at a scanning portal. The objective is to maximize the size of the powering region satisfying a particular read accuracy requirement. This research develops different methodologies and provides results for obtaining the best antenna locations to address different scenarios in supply chain applications. It addresses the case where items are static within a read portal, as well as when they might be moving on some type of material handling equipment. Various scenarios are considered for the tag orientations, including item-level applications where any orientation might be possible and case-level and pallet-level scenarios where the number of possible tag orientations might be limited.