Disaster events have emphasized the importance of healthcare response activities due to the large number of victims. For instance, Hurricane Katrina in New Orleans, in 2005, and the terrorist attacks in New York City and Washington, D.C., on September 11, 2001, left thousands of wounded people. In those disasters, although hospitals had disaster plans established for more than a decade, their plans were not efficient enough to handle the chaos produced by the hurricane and terrorist attacks. Thus, the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) suggested collaborative planning among hospitals that provide services to a contiguous geographic area during mass casualty disasters. However, the JCAHO does not specify a methodology to determine which hospitals should be included into these cooperative plans. As a result, the problem of selecting the right hospitals to include in exercises and drills at the county level is a common topic in the current preparedness stages. This study proposes an efficiency index to determine the efficient response of cooperative-networks among hospitals before an occurrence of mass casualty disaster. The index built in this research combines operations research techniques, and the prediction of this index used statistical analysis. The consecutive application of three different techniques: network optimization, data envelopment analysis (DEA), and regression analysis allowed to obtain a regression equation to predict efficiency in predefined hospital networks for mass casualty disasters. In order to apply the proposed methodology for creating an efficiency index, we selected the Orlando area, and we defined three disaster sizes. Then, we designed networks considering two perspectives, hub-hospital and hub-disaster networks. In both optimization network models the objective function pursued to: reduce the iii travel distance and the emergency department (ED) waiting time in hospitals, increase the number of services offered by hospitals in the network, and offer specialized assistance to children. The hospital network optimization generated information for 75 hospital networks in Orlando. The DEA analyzed these 75 hospital networks, or decision making units (DMU's), to estimate their comparative efficiency. Two DEAs were performed in this study. As an output variable for each DMU, the DEA-1 considered the number of survivors allocated in less than a 40 miles range. As the input variables, the DEA-1 included: (i) The number of beds available in the network; (ii) The number of hospitals available in the network; and (iii) The number of services offered by hospitals in the network. This DEA-1 allowed the assignment of an efficiency value to each of the 75 hospital networks. As output variables for each DMU, the DEA-2 considered the number of survivors allocated in less than a 40 miles range and an index for ED waiting time in the network. The input variables included in DEA-2 are (i) The number of beds available in the network; (ii) The number of hospitals available in the network; and (iii) The number of services offered by hospitals in the network. These DEA allowed the assignment of an efficiency value to each of the 75 hospital networks. This efficiency index should allow emergency planners and hospital managers to assess which hospitals should be associated in a cooperative network in order to transfer survivors. Furthermore, JCAHO could use this index to evaluate the cooperating emergency hospitals’ plans. However, DEA is a complex methodology that requires significant data gathering and handling. Thus, we studied whether a simpler regression analysis would substantially yield the same results. DEA-1 can be predicted using two regression analyses, which concluded that the average distances between hospitals and the disaster locations, and the size of the disaster iv explain the efficiency of the hospital network. DEA-2 can be predicted using three regressions, which included size of the disaster, number of hospitals, average distance, and average ED waiting time, as predictors of hospital network efficiency. The models generated for DEA-1 and DEA-2 had a mean absolute percent error (MAPE) around 10%. Thus, the indexes developed through the regression analysis make easier the estimation of the efficiency in predefined hospital networks, generating suitable predictors of the efficiency as determined by the DEA analysis. In conclusion, network optimization, DEA, and regressions analyses can be combined to create an index of efficiency to measure the performance of predefined-hospital networks in a mass casualty disaster, validating the hypothesis of this research. Although the methodology can be applied to any county or city, the regressions proposed for predicting the efficiency of hospital network estimated by DEA can be applied only if the city studied has the same characteristics of the Orlando area. These conditions include the following: (i) networks must have a rate of services lager than 0.76; (ii) the number of survivors must be less than 47% of the bed capacity EDs of the area studied; (iii) all hospitals in the network must have ED and they must be located in less than 48 miles range from the disaster sites, and (iv) EDs should not have more than 60 minutes of waiting time. The proposed methodology, in special the efficiency index, support the operational objectives of the 2012 ESF#8 for Florida State to handle risk and response capabilities conducting and participating in training and exercises to test and improve plans and procedures in the health response.

This study proposes several decision-making tools utilizing optimization and machine learning frameworks to assess and improve the safety of the workplaces. The first chapter of this study presents a novel mathematical model to optimally locate a set of detectors to minimize the expected number of casualties in a given threat area. The problem is formulated as a nonlinear binary integer programming model and then solved as a linearized branch-and-bound algorithm. Several sensitivity analyses illustrate the model's robustness and draw key managerial insights. One of the prevailing threats in the last decades, Active Shooting (AS) violence, poses a serious threat to public safety. The second chapter proposes an innovative mathematical model which captures several essential features (e.g., the capacity of the facility and individual choices, heterogeneity of individual behavioral and choice sets, restriction on choice sets depending on the location of the shooter and facility orientation, and many others) which are essential for appropriately characterizing and analyzing the response strategy for civilians under an AS exposed environment. We demonstrate the applicability of the proposed model by implementing the effectiveness of the RUN.HIDE.FIGHT.® (RHF) program in an academic environment. Given most of the past incidents took place in built environments (e.g., educational and commercial buildings), there is an urgent need to methodologically assess the safety of the buildings under an active shooter situation. Finally, the third chapter aims to bridge this knowledge gap by developing a learning technique that can be used to model the behavior of the shooter and the trapped civilians in an active shooter incident. Understanding how the civilians responded to different simulated environments, a number of actions could have been undertaken to bolster the safety measures of a given facility. Finally, this study provides a customized decision-making tool that adopts a tailored maximum entropy inverse reinforcement learning algorithm and utilizes safety measurement metrics, such as the percentage of civilians who can hide/exit in/from the system, to assess a workplace's safety under an active shooter incident.

This research contributes to gaps in the international relations literature explaining the emergence of transnational advocacy networks. Specifically, this research contributes to understanding TAN emergence due to a gap in institutional approaches to casualty recording in conflict and why actors join TANs. This TAN is particularly worthy of investigation because casualty records measure the scope of violence in a conflict and are often highly politicized and contested. Existing explanations of TAN emergence can be organized into three broad categories of analysis: sociological, political, and economic. The earliest explanations align with a sociological explanation for TANs as a mechanism for changing international norms. Social movement theorists account for TANs as a mechanism for civil society to challenge power structures. While other researchers suggest TANs should be treated like interest groups, and their emergence stems from an economic need for material incentives. This research extends the economic category of analysis and argues that actors join TANs for non-material, intangible incentives. Intangible benefits include knowledge, methodologies, data, or access to data sources. This research utilized a qualitative case study method to test all three categories of existing explanations using surveys, interviews, and archival records. Testing not only investigated hypotheses relating to the three categories of existing theories but also produced findings describing facilitators of TAN emergence, temporally-bound intangible benefits, and the types of intangible benefits available to actors. TANs are important to international politics because they influence norms, shape policies, and function as a bridge for local actors with the international community. This research produced findings with central themes about why resource-poor actors may spend their limited resources to join TANs. Further investigation into the intangible benefits available to actors joining TANs in settings other than conflict may provide greater insight into the value of intangible benefits to collective behavior. / Doctor of Philosophy / While body counts are generally presented as a measure of accountability or to raise awareness about civilian deaths in the public sphere, body counts are fiercely contested and highly politicized. This occurs during the conflict and decades after a conflict is resolved. Civilian body counts serve as political apparatuses for states and political actors to negotiate, challenge, and produce security narratives. Because of this politicization, the number of civilian casualties in violent conflict is not fully known, and their deaths' impact on the overall state's security is not well understood. While International Humanitarian Law (IHL) and human rights laws provide protection for civilians in conflict, documenting casualties is not addressed. International law does not prescribe methods for recording casualties; therefore, there is a gap in how international institutions approach accounting for casualties. In the early 2000s, facilitated by ICTs, civil society began to fill this gap by documenting casualties and collaborating across boundaries. This research traces the emergence of a Transnational Advocacy Network (TAN) that appeared in 2009 to collaborate on recording conflict casualties. This study produced five findings and contributes to understanding how ICTs facilitate TANs and identifying intangible benefits available to actors at network events that motivate their participation. Intangible benefits include knowledge, methodologies, data, or access to data sources. This research is worthy of investigation because TAN development is poorly understood yet they influence international politics by shaping norms, policies and linking local communities with international organizations.

Emergency managers are often faced with the toughest decisions that can ever be made, people's lives hang in the balance. Nevertheless, these tough decisions have to be made, and made quickly. There is usually too much information to process to make the best decisions. Decision support systems can relieve a significant amount of this onus, making decision while considering the complex interweaving of constraints and resources that define the boundary of the problem. We study these complex emergency management, approaching the problem with discrete optimization. Using our operational research knowledge to model mass casualty incidents, we seek to provide solutions and insights for the emergency managers. This dissertation proposes a novel deterministic model to optimize the casualty transportation and treatment decisions in response to a MCI. This deterministic model expands on current state of the art by; (1) including multiple dynamic resources that impact the various interconnected decisions, (2) further refining a survival function to measure expected survivors, (3) defining novel objective functions that consider competing priorities, including maximizing survivors and balancing equity, and finally (4) developing a MCI response simulation that provides insights to how optimization models could be used as decision-support mechanisms. / Doctor of Philosophy / Emergency managers are often faced with the toughest decisions that can ever be made, people's lives hang in the balance. Nevertheless, these tough decisions have to be made, and made quickly. But to make the best decisions, there is usually too much information to process. Computers and support tools can relieve a significant amount of this onus, making decision while considering the complex interweaving of constraints and resources that define the boundary of the problem. This dissertation provides a mathematical model that relates the important decisions made during a MCI response with the limited resources of the surrounding area. This mathematical model can be used to determine the best response decisions, such as where to send casualties and when to treat them. This model is also used to explore ideas of fairness and equity in casualty outcomes and examine what may lead in unfair response decisions. Finally, this dissertation uses a simulation to understand how this model could be used to not only plan the response, but also update the plan as you learn new information during the response roll-out.

This dissertation consists of three essays: essay 1, Underwriting Use of Credit Information and Firm Performance ‐ An Empirical Study of Texas Property‐Liability Insurers, essay 2, Prediction of Ratings in Property‐Liability Industry when The Organizational Form Is Endogenous, and essay 3, A Discussion of Parsimonious Methods Predicting Insurance Companies Ratings. The purpose of the first essay is to investigate the influence of underwriting use of credit information on variation in insurers’ underwriting performance. Specifically, this study addresses the following two research questions: first, what firm‐level characteristics are associated with the insurers’ decision to use credit information in underwriting? second, is there a relationship between the use of credit information and variation in insurers’ underwriting performance? The empirical results indicate that larger insurance companies, companies having more business in personal auto insurance, and those with greater use of reinsurance are more likely to use credit information in underwriting. More importantly, the results indicate that use of credit information is associated with lower variation in underwriting performance, consistent with the hypothesis that use of credit information enables insurers to better predict their losses. The purpose of the second essay is to resolve the inconsistent relationship between the organizational forms (i.e., stock versus mutual insurers) and insurers’ financial strength ratings. Specifically, this study takes into account the potential endogenous nature of organizational forms to investigate the influence of organizational forms on insurers’ financial strength ratings. The empirical results from the models employed indicate that the stock dummy variable is indeed a significant predictor of insurers’ ratings and that the relationship between the stock dummy and insurers’ financial strength ratings is not affected after the endogenous nature of organizational forms is considered. However, such relationship flips to be negative when additional rating predictors are included into the models. The purpose of the third essay is to investigate whether a logistic model is consistent in its predictions within one data set and compare the predictability and classificatory performance between the regression with a set of financial variables and the regression with principal components derived from this set of financial variables. The empirical results indicate that the models’ predictability is consistent within one data set which includes two different groups of observations. Also, the findings suggest that the principal components regression as a parsimonious model achieves the similar accuracy of estimation and fit while providing clearer interpretation of the role of the significant predictors.

Master of Science / Department of Industrial & Manufacturing Systems Engineering / Malgorzata J. Rys / Emergency department (ED) overcrowding is a nationwide problem affecting the safety and preparedness of our health care system. Many hospital EDs face significant short and intense surges in demand on a daily basis. However, the surge in demand during disaster event is not short and intense, but it is a sustained one. In order to meet this sudden surge as defined above, hospital EDs need to be more prepared and efficient to cater to increased volume of demand involving huge uncertainties. This thesis looks at the creation and use of discrete event simulation modeling using ARENA 10.0 software. In this thesis, an attempt is made to show how the different arrival patterns and time durations for which victims keep arriving affect the EDs ability to treat the patients. It is shown, how the model can be used to estimate additional resources that would be required to accommodate additional patients within the ED. Various shapes of arrival distributions were tested for different time durations. It was found that the arrival distribution with parameters (3, 4), (3, 3), (4, 2) and (2, 4) did not challenge the institutional capacity. In other words, the hospital was able to treat all the patients without compromising the quality of care up to 24 hours. However, distribution with parameter (3, 2), (2, 2), (3, 1), (1, 2), (2, 3), (2, 1), (1, 4), (1, 3), (1, 1) and (0.5, 2) did affect the system performance. Under these distributions, there was at least one patient who was either dead, LWBS or diverted. This indicates the immediacy with which victims arriving under these distributions overwhelmed the limited resources Our aim was to study, how many more resources would the ED need in order to have zero critical expire, zero Left without Being Seen (LWBS) and zero patients diverted. Arrival distribution (1, 2) was randomly selected to study this objective and it was found that for a 24 hours of simulation run time, an additional of two full trauma resources were required in order to have zero critical expire in trauma rooms area and additional of five ED beds and three nurses were required in treatment area for patients with moderate severity to have zero LWBS. With these additional resources, the ED was also able to treat all the non disaster related patients thereby having zero patients diverted. The same procedure can be used to determine the number of additional resources ED would require to treat all the victims arriving with the rest of the arrival distribution for different time periods. The simulation model built would help the emergency planners to better allocate and utilize the limited ED resources in order to treat maximum possible patients. It also helps estimate the number of additional resources that would be required in a particular scenario.

International Telemetering Conference Proceedings / October 22-25, 2001 / Riviera Hotel and Convention Center, Las Vegas, Nevada / A systems engineering development for acquisition, transmission, processing, dissemination and display of information vital to combat casualty care and related first responder activities is presented. It utilizes a synergistic combination of two existing state-of-the-art Defense Advanced Research Projects Agency/Space and Naval Warfare Systems Center San Diego (DARPA/SSCSD) technologies (Sensate-Liner and ENCOMPASS) coupled via the Enhanced Position Location Reporting System (EPLRS), an existing wireless military tactical communication data system. Transmission Security and Communication Security (TRANSEC/COMMSEC) of environmental and biomedical data is thus accomplished from the battlefield via selected data links and Ethernet. System functionality and appropriate candidate interfacing technologies will be discussed.

Since the early 1980's there has been a heightened academic interest in the field of commitment, particularly as it relates to business relationships. Knowledge of commitment continues to advance and has begun splintering and applied into specific and narrow fields. The particular field of interest in this study surrounds commitment levels in business relationships within property and casualty insurance distribution networks. The intent of understanding and enhancing commitment levels is to allow stakeholders to explore new ways to improve profitability. This can be achieved by deepening the level of understanding and knowledge of relationship partners with a view to anticipating and fulfilling their needs better than the competition. However, commitment is intangible and involves many factors including human emotion. This increases the difficulty in comprehending the whole phenomenon of commitment. To assist in furthering the knowledge in this area, transaction cost theory is examined and applied to insurance company and broker relationships. In seeking a greater understanding of the underlying drivers of commitment, this thesis investigates the theoretical contribution of transaction cost economics theory in assessing commitment levels. The purpose is to utilize the elements of transaction costs as a means to extend the awareness of how commitment is constructed, and to search for ways to improve and strengthen these relationships. The primary research method consists of three major case studies within the Canadian property and casualty insurance distribution sector. The first case study explores the perspectives of insurance brokers in Ontario. The second study reveals the perceptions of relationship managers employed with ING Canada, the country's largest property and casualty insurance company. Lastly, the research incorporates a series of interviews with ING Canada senior executives to capture their perspectives and validate the research findings from the first two case studies. These investigations into the Canadian insurance industry have provided several outputs, chief among them is the development of a conceptual model referred to as the 'Commitment Wheel'. This model has the advantages of seating affective and calculative commitment at the centre of a moving environment of commitment enablers.

當產險公司需要同時兼顧競爭力並免於破產時，適當的資產配置就是一項相當重要的決策。然而採用均數-變異數分析(mean‐variance analysis)將受到許多限制，而動態控制理論則是難以實作，因此，我們提出一個新的解決方法。這個方法主要係應用模擬最佳化的演算法，例如基礎的基因演算法(basic genetic algorithm, GA)，多階層演化策略(multi-phase evolutionary strategies, MPES)及多階層基因演算法(multi-phase genetic algorithm, MPGA)等並結合模擬模型，來求解保險公司之資產配置的問題。首先我們建立投資市場及保險業務市場的模擬模型，之後再利用本研究所發展出新的最佳化演算法來搜尋最佳的資產配置。在實務上無法實現的多期投資策略，在我們的研究架構下得以被採用，並且在比較求解結果下，多期投資策略(reallocation strategies)較定額投資策略(re‐balancing strategies)有顯著較佳的績效。在兼顧保險公司投資收益並避免破產的目標函數下，我們所提出的研究方法已證明可以用來協助保險公司建立較佳的資產配置。 / Proper asset allocations are vital for property‐casualty insurers to be competitive and remain solvent. However, popular mean‐variance analysis is limited while dynamic control theory is difficult to implement. We thus propose to apply simulation optimizations such as basic genetic algorithm (GA), multi‐phase evolutionary strategies (MPES) and multi‐phase genetic algorithm (MPGA) to the asset allocation problems of the insurers. We first construct a simulation model of the property‐casualty insurer and then develop simulation optimization techniques to search optimal investment strategies upon the simulation results. The resulted reallocation strategies perform better than re‐balancing strategies used in practice with significant margins. Therefore, our proposal researches can be used to assist insurers to construct better asset allocations.

The United States increasingly has resorted to the use of Unmanned Aerial Vehicles (UAVs) for targeted killings of terrorists as a counterterrorism strategy. More states and terrorist organizations also are acquiring UAVs and this development can lead to indiscriminate and unregulated use of UAVs. Previous researchers have indicated the surveillance ability and precise weapon delivery capacity of UAVs make them a weapon of choice for U.S. counterterrorism efforts. Although the U.S. government estimated the collateral damage involved in the use of UAVs at 3-5%, nongovernmental sources put it at 25-40%. A gap exists in the current literature regarding public perception of the use of UAVs as a counterterrorism measure and how international humanitarian law (IHL) may interpret employment of UAVs. The purpose of this quantitative, cross-sectional study is to determine if a relationship exists among public support of the use of UAVs for targeted killing, attitudes towards counterterrorism, and public perceptions of IHL. An online survey was used to collect data from 104 adult participants using the convenience sampling method. Logistic regression, ANOVA, and correlational analyses helped to determine the relationships. The outcomes contributed to the existing literature by providing important data related to public perception of the use of UAVs with the potential to enhance global peace and security. The results contributed to social change initiatives through the potential to facilitate the establishment of international and domestic legal frameworks to regulate the future employment of UAVs for targeted killing.