Search

Exploration in sequential decision problems is a computationally challenging problem. Yet, animals exhibit effective exploration strategies, discovering shortcuts and efficient routes toward rewarding sites. Characterizing this efficiency in animal exploration is an important goal in many areas of research, from ecology to psychology and neuroscience to machine learning. In this study, we aim to understand the exploration behavior of animals freely navigating a complex maze with many decision points. We propose an algorithm based on a few simple principles of animal movement from foraging studies in ecology and formalized using reinforcement learning. Our approach not only captures the search efficiency and turning biases of real animals but also uncovers longer spatial and temporal dependencies in the decisions of animals during their exploration of the maze. Through this work, we aspire to unveil a novel approach in cognitive science of drawing interdisciplinary inspiration to advancing the field's understanding of complex decision-making.

Science can be viewed as a collective, epistemic endeavor. However, a variety of factors- such as the publish-or-perish culture, institutional incentives, and publishers who favor novel and positive findings- may challenge the ability of science to accurately aggregate information about the world. Evidence of the shortcomings in the current structure of science can be seen in the replication crisis that faces psychology and other disciplines. We analyze scientific publishing through the lens of cultural evolution, framing the scientific process as a multi-generational interplay between scientists and publishers in a multi-armed bandit setting. We examine the dynamics of this model through simulations, exploring the effect that different publication policies have on the accuracy of the published scientific record. Our findings highlight the need for replications and caution against behaviors that prioritize factors uncorrelated with result accuracy.

An agent-based model for semantic search and retrieval in memory is proposed. The model seeks to generate verbal fluency lists with properties similar to those generated by humans in the semantic fluency task. This model is compared to a random walk in a semantic network in its ability to adjust to the results of 141 undergraduate students in the semantic fluency task in eight different outcomes. We found that the agent-based model fits participants' results better than the random walk model. The results were consistent with optimal foraging theories, and the distributions of the total number of words, similarities, and frequency values were similar to those generated by participants. The potential uses of this model as a virtual environment to experiment with the search and retrieval process in semantic memory are discussed.

Social rewards shape human behavior. During development, a caregiver guides a learner's behavior towards culturally aligned goals and values. How do these behaviors persist and generalize when the caregiver is no longer present, and the learner must continue autonomously? Here, we propose a model of value internalization where social feedback trains an internal social reward (ISR) model that generates internal rewards when social rewards are unavailable. Through empirical simulations, we show that an ISR model prevents agents from unlearning socialized behaviors and enables generalization in out-of-distribution tasks. Incomplete internalization, akin to "reward hacking" on the ISR, is observed when the model is undertrained. Finally, we show that our model internalizes prosocial behavior in a multi-agent environment. Our work provides a framework for understanding how humans acquire and generalize values and offers insights for aligning AI with human values.

Recent years have seen an explosion of theoretical interest, as well as increasingly fraught real-world debate, around issues to do with discourse participation. For example, marginalised groups may find themselves excluded or may exclude themselves from discourse contexts that are hostile. This not only has ethical implications, but likely impacts epistemic outcomes. The nature and scale of such outcomes remain difficult to estimate in practice. In this paper, we use agent-based modelling to explore the implications of a tendency toward `agreeableness' whereby agents might shape their communication so as to reduce direct conflict. Our simulations show that even mild tendencies to avoid disagreement can have significant consequences for information exchange and the resultant beliefs within a population.

Contemporary formal models aim to capture group polarization as the result of deliberation between rational agents. Paradigmatic models do, however, rely on rather limited agents, casting doubt on the conclusion that group polarization can be rationally reconstructed. In this paper, we use a recently developed Bayesian agent-based model of deliberation to investigate this conclusion. This model avoids problems we identify in a group of influential Bayesian polarization models. Our case study shows that a simple mechanism produces realistic patterns of group polarization: limited exchange of evidence across a sparse social network. We reflect on what our results mean for our formal understanding of rational group polarization.

We investigate the phenomenon of colexification, where a sin-gle wordform is associated with multiple meanings. Previ-ous research on colexification has primarily focused on em-pirical studies of different properties of the meanings that de-termine colexification, such as semantic similarity or meaningfrequency. Meanwhile, little attention was paid to the word-forms' properties, despite being the original approach advo-cated by Zipf. Our preregistered study examines whether wordlength influences word choice for colexification using a noveldyadic communication game (N = 64) and a computationalmodel grounded in the Rational Speech Act (RSA) framework.Contrary to initial predictions, participants did not exhibit astrong preference for efficient colexification (namely colexi-fying multiple concepts using short words, when long alter-natives are available). The results align more closely with asimpler coordination model, where dyads align on a function-ing lexical convention with relatively little influence from theefficiency of that convention. Our study highlights the pos-sibility that colexification choices are strongly determined bythe pressure for coordination, with weaker influences from se-mantic similarity or meaning frequency. This is most likelyexplained by weak pressure for efficiency in our experimentaldesign.

Social norms are often characterized as a system of rules that guide behavior. However, social norms also allow for flexibility; not entirely restricting individuals to one possible behavior. Here, we put forward an agent-based cognitive model that captures social norms as processes that are socially constructed through interactions between individuals. In this modelling work, we focus on the role of norm acquisition and conformity bias in both action production and inference-making. This computational cognitive model allows us to think about social norms along three dimensions: individual vs. collective, behavior vs. belief, and subjective vs. objective. Our simulation results show that increased conformity bias can induce misjudgments about the true desires of others and misalignment between different agents' perceptions of the social norm. However, if agents do not assume that others also conform in their behavior, this increased conformity bias does not necessarily lead to excessive misperceptions of the social norm.

In human social interactions, decisions are naturally influenced by both individual needs and the needs of others. However, it remains unclear whether cognitive robots exhibit similar needs-guided decision-making characteristics. In this study, we design a collaborative tracking task to evaluate this phenomenon. Specifically, we develop a needs-guided reinforcement learning framework that enables robots to autonomously learn and shape behavior by considering both their intrinsic needs and those of others. Our experiments highlight that the robots' inherent needs play a more crucial role in decision-making than the needs of others. In essence, our model establishes an interpretable foundation for applications in cognitive robotics.

Working memory is widely assumed to underlie multi-step planning, where representations of possible future actions and rewards are iteratively updated before determining a choice. But most working memory research focuses on a context where stimuli are presented simultaneously and the value of encoding each stimulus is independent of others. It is unclear how working memory functions in planning scenarios where the rewards of future actions unfold over time, are retained in working memory, and must be integrated for plan selection. To bridge this gap, we adapted a version of the "mouselab task" in which participants sequentially observe the reward at each node in a decision tree before selecting a plan that maximizes cumulative rewards. We specified a theoretical model to characterize the optimal encoding and maintenance strategy for this task given the working memory constraints, which trades off the cost of storing information with the potential benefit of informing later choices. The model encoded rewards in choice-relevant plans more often, in particular, rewards on the best and (to a lesser extent) worst plans. We then tested this hypotheses on human participants, who showed the same pattern in the accuracy of their explicit recall. Our study thus establishes an empirical and theoretical foundation for models of how people encode and maintain information during planning.

Human moral judgments are both precise, with clear intuitions about right and wrong, and at the same time obscure, as they seem to result from principles whose logic often escapes us. The development of Artificial Intelligence (AI) applications requires an understanding of this subtle logic if we are to embed moral considerations in artificial systems. Reinforcement Learning (RL) algorithms have emerged as a valuable interactive tool for investigating moral behavior. However, being value-based algorithms, they face difficulty when it comes to explaining deontological, non-consequentialist moral judgments. Here, in a multi-agent learning scenario based on the Producer-Scrounger Game, we show that RL agents can converge towards apparently non-consequentialist outcomes, provided the algorithm accounts for the temporal value of actions. The implications of our findings extend to integrating morality into AI agents by elucidating the interplay between learning strategies, characteristics for accounting temporal values, and methods of considering the opponent's payoff.

Theory of mind is an essential ability for complex social interaction and collaboration. Researchers in cognitive science and psychology have previously sought to integrate theory of mind capabilities into artificial intelligence (AI) agents to improve collaborative abilities (Cuzzolin, Morelli, Cirstea, & Sarahakian, 2020). We introduce the Recurrent Conditional Variational Autoencoder (RCVAE), a novel model which leverages the ability of generative models to learn rich abstracted representations of contextual behaviors to predict behavioral intent from human behavioral trajectories. Advancing on current concept learning models, this model allows for the discovery of latent intent in human behavior trajectories, while maintaining the scalability and performance of generative AI models. We show that in the Overcooked-AI environment, the RCVAE outperforms baseline Long Short-Term Memory (LSTM) models in predicting intent, achieving higher prediction accuracy and greater predictive stability. The implications of these results are significant; the RCVAE's proficiency in learning the relationship between basic actions and resulting contextual behaviors represents a significant advancement in concept learning for behavioral intent prediction.

Prior literature shows that some mechanisms, e.g., commitment, could give rise to cooperation. However, participants' diverse propensities to cooperate may limit such mechanisms' effectiveness. Thus, we bring individual differences in their propensities to cooperate into the reasoning of long-term social dynamics of cooperation through an agent-based modeling approach. Our results suggest that commitment may still guarantee cooperation when individuals have different propensities to cooperate but has weaker effects, and the setups of commitment are also important. Our study highlights the importance of integrating individual preferences in analyzing collective dynamics of a population consisting of individuals of heterogeneous characteristics, thus offering implications to facilitate cooperation in rich real-world scenarios.

Large Language models (LLM) exhibit human-like proficiency in various tasks such as translation, question answering, essay writing, and programming. Emerging research explores the use of LLMs in collective problem-solving endeavors, such as tasks where groups try to uncover clues through discussions. Although prior work has investigated individual problem-solving tasks, leveraging LLM-powered agents for group consensus and decision-making remains largely unexplored. This research addresses this gap by (1) proposing an algorithm to enable free-form conversation in groups of LLM agents, (2) creating metrics to evaluate the human-likeness of the generated dialogue and problem-solving performance, and (3) evaluating LLM agent groups against human groups using an open source dataset. Our results reveal that LLM groups outperform human groups in problem-solving tasks. LLM groups also show a greater improvement in scores after participating in free discussions. In particular, analyses indicate that LLM agent groups exhibit more disagreements, complex statements, and a propensity for positive statements compared to human groups. The results shed light on the potential of LLMs to facilitate collective reasoning and provide insight into the dynamics of group interactions involving synthetic LLM agents.

This paper presents an approach to model the internal cognitive state of decision-makers when interacting with AI to understand exchanges between agents and improve future interactions. We focus on understanding how AI suggestions are perceived by a human agent using an approach based on the technology acceptance model. The variation in the user's state is investigated when perceiving the interaction with AI by considering it as a hidden (latent) state. Using human evaluation data collected from two cases of clinical decision-making and software development scenarios, we analyse and explore the user's perceptional state during interaction. The experiment conducted employs the Bayesian belief network to represent the human perceptional model and provide a prediction of the usefulness of AI model's suggestions in the considered case. Upon introduction of cognitive states in the model, we observed an increase in predictive performance by 76–77%. Our investigation can be concluded as an attempt to identify implicit static and dynamic cognitive characteristics of users to provide personalized assistance in human-AI interaction (HAI) and collaboration in complex domains of decision-making

Artificial neural networks trained using deep-learning methods to solve a simple reference game by optimizing a task-specific utility develop efficient semantic categorization systems that trade off complexity against informativeness, much like the category systems of human languages do. But what exact type of structures in the semantic space could result in efficient categories, and how are these structures shaped by the contexts of communication? We propose a NN model that moves beyond the minimal dyadic setup and show that the emergence of convexity, a property of semantic systems that facilitates this efficiency, is dependent on the amount of variability in communication contexts across partners. We use a method of input representation based on compositional vector embeddings that is able to achieve a higher level of communication success than regular non-compositional representation methods, and can achieve a better balance between maintaining the structure of the semantic space and optimizing utility.

How does the brain control attention? The Attention Schema Theory suggests that the brain explicitly models its state of attention, termed an attention schema, for its control. However, it remains unclear under which circumstances an attention schema is computationally useful, and whether it can emerge in a learning system without hard-wiring. To address these questions, we trained a reinforcement learning agent with attention to track and catch a ball in a noisy environment. Crucially, the agent had additional resources that it could freely use. We asked under which conditions these additional resources develop an attention schema to track attention. We found that the more uncertain the agent was about the location of its attentional window, the more it benefited from these additional resources, which developed an attention schema. Together, these results indicate that an attention schema emerges in simple learning systems where attention is important and difficult to track.

Opinion averaging is a common means of judgment aggregation that is employed in the service of crowd wisdom effects. In this paper, we use simulations with agent-based models to highlight contexts in which opinion averaging leads to poor outcomes. Specifically, we illustrate the conditions under which the optimal posterior prescribed by a normative model of Bayesian argument exchange diverges from the mean belief that would be arrived at via simple averaging. The theoretical and practical implications of this are discussed.

Several recent contributions to the research on group problem solving suggest that reducing the connectivity between agents in a social network may be epistemically beneficial. This notion stems from the idea that collective problem-solving behavior may benefit from the transient diversity in agents' beliefs due to increased individual exploration and decreased social influence. At the same time, however, lower connectivity hinders the diffusion of good solutions between network members. Our simulation findings shed light on this trade-off. We identify conditions under which the less-is-more effect is likely to manifest. Our findings suggest that a community consisting of semi-isolated groups could provide an answer to the tension between diversity and diffusion.

People frequently encounter the challenge of deciding when to stop exploring options to optimize outcomes, such as when selecting an apartment in a fluctuating housing market or booking a dinner reservation on New Year's Eve. Despite experiencing these decisions on multiple occasions, people often struggle to stop searching optimally. This research investigates human learning abilities in optimal stopping tasks, focusing on feedback and knowledge of option value distributions. Through an experimental sequential choice task, we demonstrate that experience improves performance, with feedback significantly influencing learning. We also find that awareness of the value distribution reduces the duration of the search. A cognitive model accurately predicts these effects, shedding light on human learning processes.

In hierarchical perception-action (PA) learning, agents discover invariants between percepts and actions that are structured hierarchically, from very basic immediate links to higher-level, more abstract notions. In practice, existing work tends to either focus on the general theory at the expense of details of the proposed mechanisms, or specify a-priori the contents of some layers. Here, we introduce a framework that does without such constraints. We demonstrate the framework in a simple 2D environment using an agent that has minimal perceptual and action abilities. We vary the perceptual abilities of the agent to explore how the specifics of this aspect of the agent's body might affect PA learning and find unexpected consequences. The contribution of this paper is therefore twofold, (1) we add a novel framework to the literature on PA learning, using, in particular curiosity-based reinforcement learning (RL) to implement the necessary learning mechanisms, and (2) we demonstrate that even for very simple agents, the relation between the specifics of an agent's body and its cognitive abilities is not straightforward.

Conformity refers to phenomena where people match their behavior to others. Much research has focused on cases where people observe others in identical situations, saying little about its depth or generalizability. When conforming, do people revise behaviors only in that specific situation, or do they update more deeply to maintain consistent behaviors across situations? Using simulations, we first show that deep and shallow conformity leads to contrasting group dynamics; only with deep conformity can groups accumulate improvements beyond individual lifespans. We further conduct an experiment using an estimation task to examine the depths of conformity in humans. People generally extended conformist social influence to new situations without direct reference to others. However, those who simply averaged their answer with that of the direct reference showed notable failures in this generalization. Collectively, our research highlights the importance of distinguishing different depths of conformity when studying social influence and resulting group outcomes.

In this paper, we first introduce the `complex hidden profile', a previously overlooked category of hidden profiles that arises from complex inferential relations among arguments.Second, in order to investigate the conditions under which interrelated arguments can generate hidden profiles, we introduce a novel Bayesian agent-based framework for collective reasoning with complex argument structures.Finally, we show that that many possible argument structures can generate hidden profiles, even when agents do not have any information in common.

Political polarisation is growing along with its negative consequences – degradation of functional government and increases in stochastic violence. Polarisation can result from both cognitive factors affecting information processing and biased information ecosystems, but their interactions are poorly understood. We present an agent-based model combining a varyingly polarised media landscape with agents driven by homophily and uncertain (political) identity inference processes. Agents were motivated to find similar others to form an ingroup by comparing moral values expressed in response to environmentally imposed moral dilemmas. Media pushed moral values in line with either liberal or conservative values, varying in agreement and influence. Liberal agents were more satisfied (according to homophily motivations), formed larger, more stable clusters, and morally disengaged less than conservatives. Identity aligned media exposure increased liberal agents' satisfaction, but had no, or the opposite effect, on conservative agents. We conclude that media exposure asymmetrically affects political polarisation across political identities.

Various methods for social learning have been proposed within the reinforcement learning framework. These methods involve the social transmission of information within specific representational formats like policies, value, or world models. However, transmission of higher-level, model-based representations typically require costly inference (i.e., mentalizing) to ``unpack'' observable actions into putative mental states (e.g., with inverse reinforcement learning). Here, we investigate cheaper, non-mentalizing alternatives to social transmission of model-based representations that bias the statistics of experience to ``hijack'' asocial mechanisms for learning of environments. We simulate a spatial foraging task where a naïve learner learns alone or through observing a pre-trained expert. We test model-free vs. model-based learning together with simple non-mentalizing social learning strategies. Through analysis of generalization when the expert can no longer be observed and through correspondence between expert and learner representations, we show how simple social learning mechanisms can give rise to complex forms of cultural transmission.

The complexity of studying behavior and cognitive processes in realistic ecological tasks is a major challenge for cognitive scientists, behavioral ecologists, community ecologists, and the cognitive ecology community that subsumes all these fields. Here we describe a modeling approach that can be used to study the decision-making trade-offs that emerge from the coupling of nervous systems, bodies, and ecological context. To demonstrate the method, we describe an agent that must balance its need to consume resources with its need to avoid predation. We then show how to analyze the resulting behavior through the lens of behavioral trade-off schemas synthesized with neural traces measured during real-time behavior. The employment of model agents will be an important contributor to ecological theory of cognitive processes, and here we hope to convince the reader of that methodological potential.

Many recent works in the field of multi-agent reinforcementlearning via communication focus on learning what messagesto send, when to send, and whom to address such messages.Those works indicate that communication is useful for highercumulative reward or task success. However, one important limitation is that most of them ignore the importance of enforcingagents' ability to understand the received information. In thispaper, we notice that observation and communication signalsare from separate information sources. Thus, we enhance thecommunicating agents with the capability to integrate crucialinformation from different sources. Specifically, we propose amulti-modal communication method, which modulates agents'observation and communication signals as different modalitiesand performs multi-modal fusion to allow knowledge to transferacross different modalities. We evaluate the proposed methodon a diverse set of cooperative multi-agent tasks with severalstate-of-the-art algorithms. Results demonstrate the effectiveness of our method in incorporating knowledge and gaining adeeper understanding from various information sources.

Natural languages tend to minimize the linear distance between heads and their dependents in a sentence, known as dependency length minimization (DLM). Such a preference, however, has not been consistently replicated with neural agent simulations. Comparing the behavior of models with that of human learners can reveal which aspects affect the emergence of this phenomenon. This work investigates the minimal conditions that may lead neural learners to develop a DLM preference. We add three factors to the standard neural-agent language learning and communication framework to make the simulation more realistic, namely: (i) the presence of noise during listening, (ii) context-sensitivity of word use, and (iii) incremental sentence processing. While no preference appears in production, we show that the proposed factors contribute to a small but significant learning advantage of DLM for listeners of verb-initial languages. Our findings offer insights into essential elements contributing to DLM preferences in purely statistical learners.

Grief is a reaction to loss that is observed across human cultures and even in other species. While the particular expressions of grief vary significantly, universal aspects include experiences of emotional pain and frequent remembering of what was lost. Despite its prevalence, and its obvious nature, considering grief from a normative perspective is puzzling: Why do we grieve? Why is it painful? And why is it sometimes prolonged enough to be clinically impairing? Using the framework of reinforcement learning with memory replay, we offer answers to these questions and suggest, counter-intuitively, that grief may have normative value with respect to reward maximization. We additionally perform a set of simulations that identify and explore optimal grieving parameters, and use our model to account for empirical phenomena such as individual differences in human grief trajectories.

Attachment is crucial for infants' cognitive development and social relationships. Traditional attachment research has been qualitative, lacking a model to explain how infants' attachment styles develop from experience and how these are influenced by personal traits and environmental factors. We propose such a model, predicting how infants balance interaction with caregivers against exploring their surroundings. Our study is based in a grid-world environment containing an infant and caregiver agent. We vary the infant's temperamental factors (e.g., ability to regulate emotions and preferences for social vs. environmental reward), and caregiver behavior (whether positive or negative interactions are more likely). We find that different equilibria result that qualitatively correspond to different attachment styles. Our findings suggest that the characteristic exploratory behavior of each attachment style in real infants may arise from interactions of infant temperament and caregiver behaviors.

Accurately simulating human opinion dynamics is crucial for understanding a variety of societal phenomena, including polarization and the spread of misinformation. However, the agent-based models (ABMs) commonly used for such simulations often over-simplify human behavior. We propose a new approach to simulating opinion dynamics based on populations of Large Language Models (LLMs). Our findings reveal a strong inherent bias in LLM agents towards producing accurate information, leading simulated agents to consensus in line with scientific reality. This bias limits their utility for understanding resistance to consensus views on issues like climate change. After inducing confirmation bias through prompt engineering, however, we observed opinion fragmentation in line with existing agent-based modeling and opinion dynamics research. These insights highlight the promise and limitations of LLM agents in this domain and suggest a path forward: refining LLMs with real-world discourse to better simulate the evolution of human beliefs.

Generalisation of a linguistic rule can be facilitated by certain distributional characteristics. Previous work has shown that a rule is better generalised if it applies to items that (i) follow a skewed frequency distribution, or (ii) follow a uniform frequency distribution over many distinct item types. These two observations cannot be unified under explanations of rule generalisation that are based on entropy of the frequency distributions (since skewed distributions have low entropy, while a greater type count increases the entropy), nor explanations that focus on one highly-frequent type providing a basis for analogical extension (since all types in uniform distributions are equally frequent). Using an artificial language learning experiment and an agent-based model, we show that participants' generalisation behaviour is best matched by a model encoding preferential generalisation of rules containing long-tailed distributions—that is, containing a greater number of low-frequency types.

INTRODUCTION: Climate change is a global phenomenon with far-reaching consequences, and its impact on human health is a growing concern. The intricate interplay of various factors makes it challenging to accurately predict and understand the implications of climate change on human well-being. Conventional methodologies have limitations in comprehensively addressing the complexity and nonlinearity inherent in the relationships between climate change and health outcomes. OBJECTIVES: The primary objective of this paper is to develop a robust theoretical framework that can effectively analyze and interpret the intricate web of variables influencing the human health impacts of climate change. By doing so, we aim to overcome the limitations of conventional approaches and provide a more nuanced understanding of the complex relationships involved. Furthermore, we seek to explore practical applications of this theoretical framework to enhance our ability to predict, mitigate, and adapt to the diverse health challenges posed by a changing climate. METHODS: Addressing the challenges outlined in the objectives, this study introduces the Complex Adaptive Systems (CAS) framework, acknowledging its significance in capturing the nuanced dynamics of health effects linked to climate change. The research utilizes a blend of field observations, expert interviews, key informant interviews, and an extensive literature review to shape the development of the CAS framework. RESULTS AND DISCUSSION: The proposed CAS framework categorizes findings into six key sub-systems: ecological services, extreme weather, infectious diseases, food security, disaster risk management, and clinical public health. The study employs agent-based modeling, using causal loop diagrams (CLDs) tailored for each CAS sub-system. A set of identified variables is incorporated into predictive modeling to enhance the understanding of health outcomes within the CAS framework. Through a combination of theoretical development and practical application, this paper aspires to contribute valuable insights to the interdisciplinary field of climate change and health. Integrating agent-based modeling and CLDs enhances the predictive capabilities required for effective health outcome analysis in the context of climate change. CONCLUSION: This paper serves as a valuable resource for policymakers, researchers, and public health professionals by employing a CAS framework to understand and assess the complex network of health impacts associated with climate change. It offers insights into effective strategies for safeguarding human health amidst current and future climate challenges.

Agent-based modeling is a simulation method in which autonomous agents interact with their environment and one another, given a predefined set of rules. It is an integral method for modeling and simulating complex systems, such as socio-economic problems. Since agent-based models are not described by simple and concise mathematical equations, the code that generates them is typically complicated, large, and slow. Here we present Agents.jl, a Julia-based software that provides an ABM analysis platform with minimal code complexity. We compare our software with some of the most popular ABM software in other programming languages. We find that Agents.jl is not only the most performant but also the least complicated software, providing the same (and sometimes more) features as the competitors with less input required from the user. Agents.jl also integrates excellently with the entire Julia ecosystem, including interactive applications, differential equations, parameter optimization, and so on. This removes any "extensions library" requirement from Agents.jl, which is paramount in many other tools. [ABSTRACT FROM AUTHOR]

Understanding how sublethal impacts of toxicants affect population‐relevant outcomes for organisms is challenging. We tested the hypotheses that the well‐known sublethal impacts of methylmercury (MeHg) and a polychlorinated biphenyl (PCB126) would have meaningful impacts on cohort growth and survival in yellow perch (Perca flavescens) and Atlantic killifish (Fundulus heteroclitus) populations, that inclusion of model uncertainty is important for understanding the sublethal impacts of toxicants, and that a model organism (zebrafish Danio rerio) is an appropriate substitute for ecologically relevant species (yellow perch, killifish). Our simulations showed that MeHg did not have meaningful impacts on growth or survival in a simulated environment except to increase survival and growth in low mercury exposures in yellow perch and killifish. For PCB126, the high level of exposure resulted in lower survival for killifish only. Uncertainty analyses increased the variability and lowered average survival estimates across all species and toxicants, providing a more conservative estimate of risk. We demonstrate that using a model organism instead of the species of interest does not necessarily give the same results, suggesting that using zebrafish as a surrogate for yellow perch and killifish may not be appropriate for predicting contaminant impacts on larval cohort growth and survival in ecologically relevant species. Our analysis also reinforces the notion that uncertainty analyses are necessary in any modeling assessment of the impacts of toxicants on a population because it provides a more conservative, and arguably realistic, estimate of impact. Environ Toxicol Chem 2024;43:2122–2133. © 2024 SETAC [ABSTRACT FROM AUTHOR]

In this paper, we use an agent-based model (ABM) to run (counter)mobility scenarios to explore which characteristics of intermediate force capabilities (IFC) are relevant to these, and how they can affect outcomes in gray zone conflicts. Using an ABM called Map-Aware Non-Uniform Automata (MANA), developed by the New Zealand Defense Technology Agency, we implemented two scenarios where the friendly forces' mobility was limited by large groups of civilians. Then, we employed data farming and analytics methods to analyze the data and identify key parameters influencing the outcomes. The main parameters appeared to be the IFC Range, Power (a measure of the duration of the effect), and Crowd Density. Future research could include a wide range of mobility scenarios and possibly a more detailed IFC representation. [ABSTRACT FROM AUTHOR]

Previous studies conducted in the municipality of Sibaté (Colombia) have revealed alarming findings regarding asbestos exposure in the region, as it is the site of the country's first mesothelioma cluster. Non-occupational asbestos exposure events were identified in this population, and the young age of the mesothelioma cases at the time of diagnosis suggests that asbestos exposure occurred during their childhood. The creation of landfilled zones in the 1980s and 1990s, utilizing friable asbestos among other disposed materials, may have been a significant asbestos exposure event contributing to the elevated number of mesothelioma cases. The objective of this study was to model various historical exposure scenarios related to the creation and interaction of the population with asbestos-contaminated landfilled zones, in light of the absence of asbestos monitoring in the region. The models utilized a multi-agent simulation process, focusing on a 10-year period (1986–1995). Various relevant variables were incorporated into the modeling process, including, for example, the number of children playing in the landfilled zones and the percentage of children carrying asbestos fibers on their clothes to their homes. A range of values for input data for the models were utilized, spanning from very conservative numbers to exposure-promoting values. The average number of exposed individuals estimated over 750 simulation runs, considering all scenarios, was 571, with a range between 31 and 3800 exposed individuals. The use of multi-agent simulation models can assist the understanding of past asbestos exposure events, especially when there is a lack of environmental surveillance data. [ABSTRACT FROM AUTHOR]

This paper uses smart card data from Melbourne's public transport network to model and evaluate the impacts of a flexible on-demand transport system. On-demand transport is an emerging mode of urban passenger transport that relies on meeting passenger demand for travel using dynamic and flexible scheduling using shared vehicles. Initially, a simulation model was developed to replicate existing fixed-schedule bus performance and was then extended to incorporate on-demand transport services within the same network. The simulation results were used to undertake a comparative analysis which included reliability, service quality, operational efficiency, network-wide effectiveness, and environmental impacts. The results showed that on-demand buses reduced average passenger trip time by 30%, increased vehicle occupancy rates from 8% to over 50%, and reduced emissions per passenger by over 70% on an average weekday compared to fixed-schedule buses. This study also offers insights for successful on-demand transport implementation, promoting urban sustainability. It also outlines future research directions, particularly the need for accurate short-term passenger demand prediction to improve service provision and passenger experience. [ABSTRACT FROM AUTHOR]

This paper proposes the use of Pierre Bourdieu's sociological concepts of social fields, capital, and habitus to analyze water management in Colombia. By mapping the social dynamics of water management, this study examines the interactions and power relationships among agents, including government agencies, private companies, academic institutions, non-profits, and local communities. The analysis reveals how various forms of capital, such as economic, cultural, social, and symbolic, influence water management practices, policies, and the distribution of power. Integrating agent-based modeling with hydrological simulations provides a more nuanced understanding of how social dynamics influence water management. This interdisciplinary approach helps develop more adaptive and equitable strategies by capturing the complex interactions between human behavior and environmental factors. This study highlights the need to localize the analysis of the social field to capture regional customs and specific social dynamics. This localized approach ensures that water management strategies are more relevant, context sensitive, and sustainable. This paper advocates for the wider adoption of agent-based modeling in water management, proposing a methodology that combines the engineering principles of practical problem solving and adaptive design with an understanding of the social complexities in water management. [ABSTRACT FROM AUTHOR]

State-of-the-art macroeconomic agent-based models (ABMs) include an increasing level of detail in the energy sector. However, the possible financing mechanisms of renewable energy are rarely considered. In this study, an investment model for power plants is conceptualized, in which energy investors interact in an imperfect and decentralized market network for credits, deposits and project equity. Agents engage in new power plant investments either through a special purpose vehicle in a project finance (PF) structure or via standard corporate finance (CF). The model portrays the growth of new power generation capacity, taking into account technological differences and investment risks associated with the power market. Different scenarios are contrasted to investigate the influence of PF investments on the transition. Further, the effectiveness of a simple green credit easing (GCE) mechanism is discussed. The results show that varying the composition of the PF and CF strategies significantly influences the transition speed. GCE can recover the pace of the transition, even under drastic reductions in PF. The model serves as a foundational framework for more in-depth policy analysis within larger agent-based integrated assessment models. [ABSTRACT FROM AUTHOR]

This paper studies an agent‐based model of consumer demand. Agents are heterogeneous with respect to their preferences and incomes. There are two basic ingredients in the model. The first ingredient is a metric that captures the degree of heterogeneity between agents. The second ingredient is a serial computer algorithm that is used in order to compute a terminal consumption bundle at which income is exhausted and overall utility is maximized. Agents are clustered into heterogeneous groups based on their preferences and incomes. We extract information about the evolution of consumer expenditure under different price regimes and the buildup of optimal demand for varying levels of income and preference parameter values. These features cannot be obtained in the classical framework of static utility maximization. Our agent‐based data‐driven methodology can be applied to any relevant data set and so provide a reliable model for forecasting demand given some agent characteristics. [ABSTRACT FROM AUTHOR]

This work aims to develop and validate a framework for the multiscale simulation of the biological response to ionizing radiation in a population of cells forming a tissue. We present TOPAS-Tissue, a framework to allow coupling two Monte Carlo (MC) codes: TOPAS with the TOPAS-nBio extension, capable of handling the track-structure simulation and subsequent chemistry, and CompuCell3D, an agent-based model simulator for biological and environmental behavior of a population of cells. We verified the implementation by simulating the experimental conditions for a clonogenic survival assay of a 2-D PC-3 cell culture model (10 cells in 10,000 µm2) irradiated by MV X-rays at several absorbed dose values from 0–8 Gy. The simulation considered cell growth and division, irradiation, DSB induction, DNA repair, and cellular response. The survival was obtained by counting the number of colonies, defined as a surviving primary (or seeded) cell with progeny, at 2.7 simulated days after irradiation. DNA repair was simulated with an MC implementation of the two-lesion kinetic model and the cell response with a p53 protein-pulse model. The simulated survival curve followed the theoretical linear–quadratic response with dose. The fitted coefficients α = 0.280 ± 0.025/Gy and β = 0.042 ± 0.006/Gy2 agreed with published experimental data within two standard deviations. TOPAS-Tissue extends previous works by simulating in an end-to-end way the effects of radiation in a cell population, from irradiation and DNA damage leading to the cell fate. In conclusion, TOPAS-Tissue offers an extensible all-in-one simulation framework that successfully couples Compucell3D and TOPAS for multiscale simulation of the biological response to radiation. [ABSTRACT FROM AUTHOR]

The digital twin (DT) mainly acts as a virtual exemplification of a real-world entity, system, or process via multiphysical and logical models, allowing the capture and synchronization of its functions and attributes. The bridge between the actual system and the digital realm can be utilized to optimize the system's performance, and forecast and predict its behavior. Incorporating intelligent and adaptive reasoning mechanisms into DTs is crucial to enable them to reason, adapt, and take efficacious actions in complex and dynamic environments. To this end, we introduce an approach for deploying agent-based DTs for cyber-physical systems. The foundation pillars of this approach are (1) integrating the concepts, entities, and relations of Zeigler's modeling and simulation framework from the perspective of agent-based DTs; (2) utilizing an expandable and scalable architecture for designing and materializing these twins, which handily enables extending and scaling physical and digital assets of the system; and finally (3) a two-tier reasoning strategy; reactive and rational models are conceptually redefined in the context of the modeling and simulation framework of agent-based DTs and technically deployed to boost the adaptive reasoning and decision-making function of DTs. As a result, an implemented simulation and control platform for a multi-robot system demonstrates the approach's applicability and feasibility, manifesting its usability and efficiency. The platform represents physical entities as autonomous agents, creates their DTs, and assigns adequate reasoning capability to promote adaptive planning, autonomous resource management, and flexible logical decision-making to handle different situations and scenarios. [ABSTRACT FROM AUTHOR]

This study presents extended Immunity Agent-Based Model (IABM) simulations to evaluate vaccination strategies in controlling the spread of infectious diseases. The application of IABM in the analysis of vaccination configurations is innovative, as a vaccinated individual can be infected depending on how their immune system acts against the invading pathogen, without a pre-established infection rate. Analysis at the microscopic level demonstrates the impact of vaccination on individual immune responses and infection outcomes, providing a more realistic representation of how the humoral response caused by vaccination affects the individual's immune defense. At the macroscopic level, the effects of different population-wide vaccination strategies are explored, including random vaccination, targeted vaccination of specific demographic groups, and spatially focused vaccination. The results indicate that increased vaccination rates are correlated with decreased infection and mortality rates, highlighting the importance of achieving herd immunity. Furthermore, strategies focused on vulnerable populations or densely populated regions prove to be more effective in reducing disease transmission compared to randomly distributed vaccination. The results presented in this work show that vaccination strategies focused on highly crowded regions are more efficient in controlling epidemics and outbreaks. Results suggest that applying vaccination only in the densest region resulted in the suppression of infection in that region, with less intense viral spread in areas with lower population densities. Strategies focused on specific regions, in addition to being more efficient in reducing the number of infected and dead people, reduce costs related to transportation, storage, and distribution of doses compared to the random vaccination strategy. Considering that, despite scientific efforts to consolidate the use of mass vaccination, the accessibility, affordability, and acceptability of vaccines are problems that persist, investing in the study of strategies that mitigate such issues is crucial in the development and application of government policies that make immunization systems more efficient and robust. [ABSTRACT FROM AUTHOR]

Despite interest in preindustrial markets, archaeological discussions have largely been limited to proposing methods to determine the presence or absence of market exchange in ancient societies. While these contributions are important, methodological limitations have prevented theoretical considerations of the emergence and evolution of marketplaces and market exchange in prehistory. We propose that agent-based modeling provides a window to explore physical conditions and agent behaviors that facilitate the emergence of customary exchange locations and how such locations may evolve into socially embedded institutions. The model we designed suggests that simple bartering rules among agents can generate concentrated locations of exchange and that spatial heterogeneity of resources is the most important factor in facilitating the emergence of such locales. Furthermore, partner-search behaviors and exchange of information play a key role in the institutionalization of the marketplace. The results of our simulation suggest that marketplaces can develop, even with the absence of formalized currency or central planning, as a consequence of collective strategies taken up by agents to reduce exchange partner-search costs and make transactions more frequent and predictable. The model also suggests that, once established as a social institution, marketplaces may become highly conservative and resistant to change. As such, it is inferred that bottom-up and/or top-down interventions may have often been required to establish new marketplaces or relocate marketplaces to incorporate new resources, resolve supply–demand imbalances, or minimize rising economic costs that arise as a result of social, political, and economic change. [ABSTRACT FROM AUTHOR]

Effective disaster management can help mitigate both human and financial damage. However, the absence of appropriate disaster preparedness and management efforts can increase the vulnerability of international tourists to natural disasters. International tourists differ from residents and local tourists in various ways, including limited knowledge of disasters and disaster responses, use of different languages, limited access to signs, and shelter locations and evacuation routes. This study developed an agent-based evacuation model to determine the differences in evacuation behaviors among international tourists, residents, and local tourists during a tsunami. Further, it clarified the impact of both soft and hard countermeasures such as information provision and shelter capacity expansion. Our case study was conducted in the Minato Bay area in Osaka, Japan. The simulation results revealed disparities in the evacuation behaviors between international tourists and residents in the arrival times at shelters. The enhanced sign accessibility and provision of signs in multiple languages significantly reduced the arrival time of international tourists at shelters. Consequently, it is necessary to improve disaster management plans that ensure information provision in multiple languages and the establishment of temporary shelters near tourism spots to support international tourists during disaster evacuation. • Disparities exist in the evacuation behaviors of international tourists and residents. • International tourists following other evacuees take more time to reach shelters. • Enhanced sign accessibility reduces international tourists' arrival times. • Information provision in multiple languages reduces international tourists' arrival times. [ABSTRACT FROM AUTHOR]

This research examines the use of Shared Autonomous Electric Vehicles (SAEVs) in 374 U.S. small and medium-sized urban areas, focusing on fleet and infrastructure needs through agent-based simulations. It assesses metrics such as fleet size, trips per vehicle, and charging station requirements, considering two charger types: Level 2 and Level 3. The findings show significant spatial differences in SAEV operations and infrastructure across these cities. Statistical analysis links these variations to regional road networks and travel patterns. The study finds Level 3 chargers more efficient, requiring fewer stations and enabling more trips per vehicle compared to Level 2 chargers. Furthermore, Level 3 chargers exhibit a greater number of trips per SAEV and a higher ratio of vehicles to charging stations. These findings highlight the significance of considering charging infrastructure characteristics to optimize SAEV fleet performance and promote sustainable transportation systems in urban areas. This study significantly contributes by identifying the spatial variation and correlates of the SAEVs' operational and charging infrastructural performance. Policymakers, urban planners, and transportation service providers can leverage these insights to design and implement effective charging infrastructure for SAEV fleets, thereby advancing the transition to cleaner and more efficient mobility solutions. • This study investigates the potential of Shared Autonomous Electric Vehicles (SAEVs). • Agent-based simulations examined critical metrics such as fleet size and charging station needs. • The analysis includes two types of charging stations, Level 2 and Level 3. • Notable spatial variations in SAEV fleet and charging infrastructure performance are identified. • Regional road network characteristics and travel demand patterns contribute to spatial variations. [ABSTRACT FROM AUTHOR]