A growing body of research uses computational models to study political decision making and behavior such as voter turnout, vote choice, party competition, social networks, and cooperation in social dilemmas. Advances in the computational modeling of political decision making are closely related to the idea of bounded rationality. In effect, models of full rationality can usually be analyzed by hand, but models of bounded rationality are complex and require computer-assisted analysis. Most computational models used in the literature are agent based, that is, they specify how decisions are made by autonomous, interacting computational objects called “agents.” However, an important distinction can be made between two classes of models based on the approaches they take: behavioral and information processing. Behavioral models specify relatively simple behavioral rules to relax the standard rationality assumption and investigate the system-level consequences of these rules in conjunction with deductive, game-theoretic analysis. In contrast, information-processing models specify the underlying information processes of decision making—the way political actors receive, store, retrieve, and use information to make judgment and choice—within the structural constraints on human cognition, and examine whether and how these processes produce the observed behavior in question at the individual or aggregate level. Compared to behavioral models, information-processing computational models are relatively rare, new to political scientists, and underexplored. However, focusing on the underlying mental processes of decision making that must occur within the structural constraints on human cognition, they have the potential to provide a more general, psychologically realistic account for political decision making and behavior.
Agent-based computational modeling (ABM, for short) is a formal and supplementary methodological approach used in international relations (IR) theory and research, based on the general ABM paradigm and computational methodology as applied to IR phenomena. ABM of such phenomena varies according to three fundamental dimensions: scale of organization—spanning foreign policy, international relations, regional systems, and global politics—as well as by geospatial and temporal scales. ABM is part of the broader complexity science paradigm, although ABMs can also be applied without complexity concepts. There have been scores of peer-reviewed publications using ABM to develop IR theory in recent years, based on earlier pioneering work in computational IR that originated in the 1960s that was pre-agent based. Main areas of theory and research using ABM in IR theory include dynamics of polity formation (politogenesis), foreign policy decision making, conflict dynamics, transnational terrorism, and environment impacts such as climate change. Enduring challenges for ABM in IR theory include learning the applicable ABM methodology itself, publishing sufficiently complete models, accumulation of knowledge, evolving new standards and methodology, and the special demands of interdisciplinary research, among others. Besides further development of main themes identified thus far, future research directions include ABM applied to IR in political interaction domains of space and cyber; new integrated models of IR dynamics across domains of land, sea, air, space, and cyber; and world order and long-range models.
Louise K. Comfort
Managing critical infrastructures presents a specific set of challenges to crisis managers. These systems include electrical power; communications; transportation; and water, wastewater, and gas line distribution systems. Such infrastructures undergird the continued operation of communities in a modern society. Designed for efficiency, these technical systems operate interdependently, which makes them vulnerable to the stress of extreme events. Changes in population, demographics, land use, and economic and social conditions of communities exposed to hazards have significantly increased the number of people dependent on critical infrastructures in regions at risk. Although advances in science, technology, and engineering have introduced new possibilities for the redesign, maintenance, and retrofit of built infrastructure to withstand extreme events, the complexity of the task has exceeded the capacity of most public and private agencies to anticipate the potential risk and make investments needed to upgrade infrastructures before damage occurs. A mix of public and private ownership of infrastructure systems further complicates the task of ensuring public safety and security in crisis. Public agencies cannot protect communities alone. FEMA has developed a “whole of nation” approach to strengthen cross-jurisdictional linkages with state, county, and municipal emergency managers as well as private and nonprofit organizations. Computational modeling facilitates the exploration of alternative approaches to managing risk generated among a range of actors, interdependent infrastructures, and types of hazard events. Advanced uses of sensors, telemetry, and graphic display of changing performance for critical infrastructure provide timely, accurate information to reduce uncertainty in crisis events. Such technologies enable crisis managers to track more accurately the impact of extreme events on the populations and infrastructures of communities at risk, and to anticipate more reliably the likely consequences of future hazardous events. A basic shift has occurred in the assessment of risk. The focus is no longer on calculating the damage from past events, but on anticipating and reducing the consequences of future hazards, based on sound, scientific evidence as well as local experience and knowledge. Recognizing communities as complex, adaptive systems, crisis managers strive to create a continual learning process that enables residents to monitor their changing environment, use systematically collected data as the basis for analysis and change, and modify policies and practice based on valid evidence from actual environments at risk. Visualization constitutes a key component of collective learning. In complex settings, people comprehend visual images more readily than written or aural directions. Using graphic technologies to display emerging risk at multiple levels simultaneously provides an effective means to guide particular decisions at intermediate (meso) and local levels of operation. For communities seeking to reduce risk, investment in information technologies to enable rapid, community-wide access to interactive communication constitutes a major step toward building capacity not only for managing risk to critical infrastructure but also in maintaining continuity of operations for the whole community in extreme events.