abstract: This dissertation consists of three essays, each examining distinct aspects about public organization adaptation to extreme events using evidence from public transit agencies under the influence of extreme weather in the United States (U.S.). The first essay focuses on predicting organizational adaptive behavior. Building on extant theories on adaptation and organizational learning, it develops a theoretical framework to uncover the pathways through which extreme events impact public organizations and identify the key learning mechanisms involved in adaptation. Using a structural equation model on data from a 2016 national survey, the study highlights the critical role of risk perception to translate signals from the external environment to organizational adaptive behavior. The second essay expands on the first one to incorporate the organizational environment and model the adaptive system. Combining an agent-based model and qualitative interviews with key decision makers, the study investigates how adaptation occurs over time in multiplex contexts consisting of the natural hazards, organizations, institutions and social networks. The study ends with a series of refined propositions about the mechanisms involved in public organization adaptation. Specifically, the analysis suggests that risk perception needs to be examined relative to risk tolerance to determine organizational motivation to adapt, and underscore the criticality of coupling between the motivation and opportunities to enable adaptation. The results further show that the coupling can be enhanced through lowering organizational risk perception decay or synchronizing opportunities with extreme event occurrences to promote adaptation. The third essay shifts the gaze from adaptation mechanisms to organizational outcomes. It uses a stochastic frontier analysis to quantify the impacts of extreme events on public organization performance and, importantly, the role of organizational adaptive capacity in moderating the impacts. The findings confirm that extreme events negatively affect organizational performance and that organizations with higher adaptive capacity are more able to mitigate those effects, thereby lending support to research efforts in the first two essays dedicated to identifying preconditions and mechanisms involved in the adaptation process. Taken together, this dissertation comprehensively advances understanding about public organization adaptation to extreme events. / Dissertation/Thesis / Doctoral Dissertation Public Administration and Policy 2020

No description available.

i Abstract In this work, we describe a novel methodology to analyse the price elasticity of demand. This method combines an artificial neural network that serves as the model of the behaviour of the customers and a subsequent simulation based on this model. We present the validation of our approach using a real-world dataset obtained from an e-commerce retailer and demonstrate its advantages, notably the ability to estimate the elasticity in distinct price points and the inclusion of the complete pricing situations (not only product's own price). JEL Classification C45, C44, C15, D12 Keywords price elasticity of demand, artificial neural net- work, agent-based model Title A simulation based analysis of price elasticity of demand Author's e-mail Supervisor's e-mail

<p>In many economic settings agents behave strategically. Understanding and, sometimes regulating, that behavior is often crucial to enhance the efficiency with which scarce resources are allocated. A peculiar feature of economics is that cooperation among agents sometimes boosts efficiency, and sometimes hinders it. Social dilemmas, highly ubiquitous in economics, are situations in which cooperation boosts efficiency. Highly concentrated markets where a few firms operate, are situations in which cooperation (also known as collusion) among firms hinders efficiency. In such markets competition, rather than cooperation, boosts efficiency. In this dissertation, I study how uncertainty affects cooperation in social dilemmas, and how the presence of cooperative firms affects competition in concentrated markets.</p> <p><br></p> <p>Both of the settings I study in this dissertation (social dilemmas with noisy payoffs and duopsony with endogenous location and pricing strategy) face a similar challenge. Their complexity compromises the tractability of conventional equilibrium concepts. In other words, Nash equilibria do not exist, or there is a multiplicity of equilibria. This, in turn, precludes comparative static analyses characterizing the effect of exogenous market forces (uncertainty and firm ownership structure) on market and welfare outcomes.</p> <p><br></p> <p>I address this key challenge through a combination of genetic algorithms and laboratory experiments. A genetic algorithm consists of a selection process that identifies strategies that perform better than others, on average. Therefore, surviving strategies constitute, in a sense, average best responses. More than one strategy may survive. This happens when none of the surviving strategies is weakly dominated by the other surviving strategies. An equilibrium is a combination of surviving strategies. In this context, a comparative static analysis consists of the change in equilibrium (combination of surviving strategies) due to a change in exogenous forces. These comparative static analyses generate testable hypotheses. In Essays 1 and 2, I implement laboratory experiments to test these hypotheses.</p> <p><br></p> <p>In Essay 1, I compare infinitely repeated social dilemmas with deterministic and noisy payoffs. I test whether noise in payoffs (where noisy payoffs are generated by a random shock and are uncorrelated amongst agents), which introduces imperfect monitoring, affects cooperation. Experimental evidence shows that imperfect monitoring reduces cooperation because it hinders agents’ ability to threaten defectors with a reciprocal defection. Therefore, noise reduces efficiency by unraveling cooperation in social dilemmas. In Essay 2, I study whether correlation among agents’ noisy payoffs strengthens monitoring and restores cooperation. Experimental evidence shows that stronger (though still imperfect) monitoring due to correlation helps cooperation if and only if agents are prone to cooperate in the initial rounds of the repeated game. Therefore, correlation among shocks affecting agents’ payoffs may or may not increase efficiency depending on the type of players participating in the social dilemma.</p> <p><br></p> <p>Finally, in Essay 3, I use a genetic algorithm to generate comparative statics characterizing the effect of a cooperative firm on market equilibrium and efficiency in a spatial duoposony. A Nash equilibrium in this setting does not exist when location, price, and the degree of spatial price discrimination are all endogenous in the seminal Hotelling’s model. I use a genetic algorithm to identify a stable equilibrium in this setting. I find that a cooperative firm increases efficiency. But, counterintuitively, it does so when the cooperative does not directly compete with the privately owned firm. This is because the cooperative maximizes market share when its procurement region does not overlap with the privately owned firm’s procurement region.</p> <p><br></p>

Questa tesi è composta da due saggi indipendenti. Il saggio presentato nel Capitolo 1 studia la conformità fiscale all'interno di un modello basato su agenti. Il modello è progettato tenendo conto di una serie di regole fiscali in vigore in Italia e calibrato con micro-dati forniti dall'autorità fiscale italiana. I risultati delle simulazioni mostrano che, considerando livelli di deterrenza realistici, agenti strettamente razionali generano un livello (medio) di non conformità fiscale sostanzialmente superiore a quello suggerito dai dati empirici. Quando includiamo nel processo decisionale dell’agente il calcolo e l’aggiornamento della probabilità soggettiva di subire un controllo, così come l’attitudine alla conformità sociale e gli effetti di rete, il modello fornisce risultati maggiormente in linea con l'evidenza empirica. Il saggio presentato nel Capitolo 2 impiega diverse tecniche di apprendimento automatico, con l'obiettivo di identificare quei contribuenti che hanno maggiore probabilità di aumentare l’importo della loro dichiarazione dei redditi dopo essere stati controllati dall'autorità fiscale. Tra i metodi impiegati, la foresta casuale ha garantito la maggiore accuratezza predittiva. Per valutare l'utilità pratica del nostro approccio, calcoliamo l'aumento del reddito netto riportato dai contribuenti identificati dal modello random forest. Troviamo che, in media, questo aumento è significativo rispetto alla media di tutti i contribuenti ispezionati. Riteniamo, dunque, che il nostro approccio possa rivelarsi uno strumento utile al fine di individuare e selezionare quei contribuenti che hanno una maggiore probabilità di dichiarare un reddito più alto in seguito ad un controllo, consentendo, quindi, una migliore allocazione delle - tipicamente scarse - risorse finanziarie a disposizione dell’autorità fiscale nell'ambito della sua attività ordinaria di controllo. / The essay presented in Chapter 1 studies tax compliance within an agent-based framework. The model is designed according to a set of normative taxing rules for the Italian case and calibrated with micro-data provided by the Italian tax authority. Simulation results show that, under realistic deterrence levels, strict rational agents generate a (average) level of tax noncompliance substantially higher than that suggested by the empirical data. When subjective audit probability computing and updating as well as social conformity attitude and network effects are included in the decision process, the model provides results more in line with the empirical evidence. The essay presented in Chapter 2 employs several machine learning techniques, with the aim to identify those taxpayers who are more likely to increase their net income declarations after being audited by the tax authority. Among the employed methods, random forest guaranteed higher predictive accuracy. In order to assess the practical utility of our approach, we compute the reported net income increase by taxpayers identified through the random forest model. We find that, on average, this increase is significant compared to the average of all the inspected taxpayers. We believe that our approach could prove a useful tool in order to identify and select those taxpayers who are more likely to increase the income reporting after an audit, therefore allowing for a better allocation of the – typically scarce – financial resources available to the tax authority for its ordinary auditing activities.

Huge amount of medical data has been generated in practical experiments which makes data analysis a challenging problem. This requires novel techniques to be developed. The improvements in computational power suggest to use computerbased modeling approaches to process a large set of data. One of the important systems in the human body to be investigated is the immune system. The previous studies of medical scientists and ongoing experiments at Karolinska Institute provide information about the human immune system. This information includes attributes of human immune system’s blood cells and the interactions between these cells. This interactions are provided as ‘if-then’ logical rules. Each rule verifies a condition on the attribute of one cell and it may initiate interaction processes to modify the attributes of other cells. A specific temporal value is associated to each process to quantify the speed of that process in the body (i.e., slow, medium, fast). We propose an agent-based model (ABM) to study human immune system cells and their interactions. The ABM is selected to overcome the complexity of large amount of data and find emergent properties and behavior patterns of the cells. Immune system cells are modeled as autonomous agents which have interactions with each other. Different values of a cell attributes define possible states of the cell and the collection of states of all cells constructs the state of the whole agent-based model. In order to consider the state transitions of the cells, we used a finite state machine (FSM). The first state is constructed from the input initial values for the cells and considering the logical time of 1. In each step, the program goes one time unit further and computes next state by applying the changes based on the cells’ interactions rules. This evolution of states in time is similar to game of life (GOL) automaton. The final model based on three modeling approaches of ABM, FSM and GOL are used to test medical hypothesis related to human immune system. This model provides a useful framework for medical scientists to do experiments on the cells’ attributes and their interaction rules. Considering a set of cells and their interactions, the proposed framework shows emergent properties and behavior patterns of the human immune system.

Common pool resource management systems are complex to manage due to the absence of a clear understanding of the effects of users’ behavioral characteristics. Non-cooperative decision making based on individual rationality (as opposed to group rationality) and a tendency to free ride due to lack of trust and information about other users’ behavior creates externalities and can lead to tragedy of the commons without intervention by a regulator. Nevertheless, even regulatory institutions often fail to sustain natural common pool resources in the absence of clear understanding of the responses of multiple heterogeneous decision makers to different regulation schemes. While modeling can help with our understanding of complex coupled human-natural systems, past research has not been able to realistically simulate these systems for two major limitations: 1) lack of computational capacity and proper mathematical models for solving distributed systems with self-optimizing agents; and 2) lack of enough information about users’ characteristics in common pool resource systems due to absence of reliable monitoring information. Recently, different studies have tried to address the first limitation by developing agent-based models, which can be appropriately handled with today’s computational capacity. While these models are more realistic than the social planner’s models which have been traditionally used in the field, they normally rely on different heuristics for characterizing users’ behavior and incorporating heterogeneity. This work is a step-forward in addressing the second limitation, suggesting an efficient method for collecting information on diverse behavioral characteristics of real agents for incorporation in distributed agent-based models. Gaming in interactive virtual environments is suggested as a reliable method for understanding different variables that promote sustainable resource use through observation of decision making and iii behavior of the resource system beneficiaries under various institutional frameworks and policies. A review of educational or "serious" games for environmental management was undertaken to determine an appropriate game for collecting information on real-agents and also to investigate the state of environmental management games and their potential as an educational tool. A web-based groundwater sharing simulation game—Irrigania—was selected to analyze the behavior of real agents under different common pool resource management institutions. Participants included graduate and undergraduate students from the University of Central Florida and Lund University. Information was collected on participants’ resource use, behavior and mindset under different institutional settings through observation and discussion with participants. Preliminary use of water resources gaming suggests communication, cooperation, information disclosure, trust, credibility and social learning between beneficiaries as factors promoting a shift towards sustainable resource use. Additionally, Irrigania was determined to be an effective tool for complementing traditional lecture-based teaching of complex concepts related to sustainable natural resource management. The different behavioral groups identified in the study can be used for improved simulation of multi-agent groundwater management systems.

Decision-making by one or more individuals to select a course of action is predicated on the values and preferences to identify, choose options, and finally select the option that is evaluated to be the “best option.” Decision theory provides the means to model and analyze both the processes and options available to the decision-makers. This dissertation assembled in three phases: 1) An effort to collect and review existing literature relating to the concept of expanding decision analysis options to provide a model of decision-making made with time-dependent factors along with uncertainty and risk. Further, adding the concept of a decision to update time-dependent decision data in a Bayesian fashion aids in modeling decision-making thought processes. This review included a total of 395 research artifacts. 2) Development of a technical approach using the information gathered in the literature review to guide planning for a decision-making simulation of individuals and organizations. The approach emphasizes creating a decision-making simulation framework with capabilities to model time-dependent factors, information processing and communication, and fuzzy-stochastic data. 3) Use of the technical approach to develop a simulation framework to simulate complex decision-making and work packages at multiple levels in an organization using time-dependent factors, information processing and communication, and fuzzy-stochastic data. Using a Discrete Event Simulation (DES) and Agent-Based Models (ABM) to simulate people and their interactions, this framework was then be used to simulate decision-making and work processes within an organization. Ultimately, the Agile Enterprise Simulation (AES) capability was created and demonstrated.

Whitebark pine is an important keystone and foundation species in western North American mountain ranges, and facilitates tree island development in Rocky Mountain treelines. The manifestation of white pine blister rust in the cold and dry treelines of the Rockies, and the subsequent infection and mortality of whitebark pines raises questions as to how these extreme environments harbor the invasive disease, and what the consequences may be for treeline dynamics. This dissertation research comprises three studies that investigate abiotic factors influential for blister rust infection in treeline whitebark pines, how disease coupled with changing climate may affect whitebark pine treeline dynamics, and the connection between treeline spatial patterns and disease. The first study examined the spatial variation of blister rust infection in two whitebark pine treeline communities, and potential topographic correlates. Using geospatial and field approaches to generate high resolution terrain models of treeline landscapes, microtopography associated with solar radiation and moisture were found most influential to blister rust infection in treeline whitebark pines. Using field-based observations from sampled treeline communities, the second study developed an agent-based model to examine the effects of disease and climate on treeline pattern and process. Treeline dynamics were simulated under five hypothetical scenarios to assess changes in tree spatial patterns and populations. Blister rust-induced loss of whitebark pines resulted in a decline of facilitative processes, and an overall negative treeline response to disease—despite the beneficial effects of climate amelioration. The objective of the third study was to explore whether spatial patterns of tree proximity, size, and growth infer disease patterns. Comparing spatial patterns of tree characteristics between diseased and undiseased treeline communities, I found that trees growing near trees with larger stem diameters, and larger tree islands, tended to have more blister rust cankers, and displayed clustered spatial patterns. Undiseased treeline patterns revealed near neighbors smaller in stem diameter and tree island size, and were randomly dispersed. Blister rust diseased whitebark pines reveal spatial autocorrelation, despite the complex blister rust disease life cycle. Overall, findings from this dissertation reveal the implications of invasive disease on sensitive treeline ecotones dependent on a keystone species. / Ph. D.

Agent-based models are computer models made up of agents that can exist in a finite number of states. The state of the system at any given time is determined by rules governing agents' interaction. The rules may be deterministic or stochastic. Optimization is the process of finding a solution that optimizes some value that is determined by simulating the model. Optimal control of an agent-based model is the process of determining a sequence of control inputs to the model that steer the system to a desired state in the most efficient way. In large and complex models, the number of possible control inputs is too large to be enumerated by computers; hence methods must be developed for use with these models in order to find solutions without searching the entire solution space. Heuristic algorithms have been applied to such models with some success. Such algorithms are discussed; case studies of examples from biology are presented. The lack of a standard format for agent-based models is a major issue facing the study of agent-based models; presentation as polynomial dynamical systems is presented as a viable option. Algorithms are adapted and presented for use in this framework. / Master of Science