In today’s high-tech age, the advancement of information and communications technology has played a significant role in solving some of the real-world problems in our societies. However, due to the nondeterministic behavioral patterns of human being...
In today’s high-tech age, the advancement of information and communications technology has played a significant role in solving some of the real-world problems in our societies. However, due to the nondeterministic behavioral patterns of human beings, the management of its complex dynamics remains a challenge. In particular, two factors hinder policymakers and researchers from applying various data-driven approaches to the management of social systems: i) many modern societies lack the long history and stable indices required for the reproduction of past trends, and ii) it is always time-consuming and expensive to accumulate sufficient data for complete analyses. To overcome these obstacles, specialists in modeling and simulation (M&S) recommended one practical approach, i.e., constructing a virtual model similar to the target system and analyzing its simulation outputs depending on the available information in a given situation. Agent-based modeling and simulation (ABMS) is an M&S technique that assembles autonomous and interacting decision makers (e.g., human beings, groups, and institutions) in an effort to simulate their collective behaviors. Recently, this technique has attracted attention as a new M&S paradigm, and its inconclusive theorization in representing human systems has disrupted the practical dissemination of ABMS for real-world problems.
The main goal of this study was to investigate a better framework for agent-based models that describe how people actually behave across interdisciplinary domains. As a step toward this far-reaching ambition, this study attempts to develop a modeling formalism that enables researchers to explore a truer agent-based representation of multidisciplinary human systems grounded in systems theory and ecological findings relating to human behavior. The target system is specified as the social and biophysical environments in which availability-based imitation and movement become manifested according to the affordance theory. The research scope is limited to illustrations of two fundamental actions of human beings in three different application domains, namely, pedestrian dynamics (in physical space), nuptiality dynamics (in social space), and evacuation dynamics (in integrated space).
To this end, the study is composed of three parts: i) a literature review pertaining to modeling formalisms for human systems in the context of ABMS, ii) the development of a formalism based on the latest findings in M&S engineering and cognitive sciences, and iii) three illustrations of the developed formalism and plausibility tests of the simulated dynamics. The first segment presents the research background for understanding the full details of the proposed formalism, including the concept of ABMS, its theoretical trends in representing human behaviors, affordance theory, and formalisms for multicomponent systems. In summary, well-cited studies pertaining to the concepts, structures, and terminologies of ABMS are surveyed for general scholarly sentiments regarding them. In particular, the review also highlights the latest findings in the theory of affordances and formalisms, which have been incorporated in predecessors to this research. The second segment explains the proposed formalism through two different perspectives based on the aforementioned theoretical foundations. It presents not only a conceptual framework for cognitive science and its related disciplines but also a mathematical template based on DEVS for M&S engineers. Subsequently, an illustration of pedestrian dynamics is implemented within an off-the-shelf M&S software (i.e., AnylogicTM) to enable easier reproductions and verify a default working mechanism of the proposed formalism. The last segment includes case studies on two different domains, namely, nuptiality and evacuation dynamics. The first section of this part develops an agent-based approach for exploring nuptiality dynamics and marriage heuristics in small-world networks. The other section demonstrates evacuation in a radiological emergency based on the fundamental social and physical behaviors (i.e., movement and imitation) of human beings. Plausibility test of the simulated output has also been conducted at the level of a proof-of-concept due to limited empirical evidence within the target system. At the beginning of each section, the research contexts of both applications are specified to clearly differentiate the proposed approaches.
The key novelty of this study is the incorporation of ecological findings on ubiquitous human behaviors into the latest modeling formalisms, which are specialized in ABMS, to fill the gap between the perspectives of cognitive scientists and M&S engineers. The developed formalism focuses on human nature in terms of accessible information rather than perfectly taking into account the entirety of a situation. It enables agent-based modelers to more clearly simulate the complex dynamics driven by the bounded rationality of human beings. The formalism could support the provision of an extensible digital twin for factories, cities, etc., in which multidisciplinary issues coexist. In conjunction with cyber-physical systems (CPS), human-in-the-loop (HITL), and other artificial intelligence (AI) technologies, it is anticipated for the proposed agent-based approach to contribute to a more convenient society for humans relying on automatic systems within themselves.