Large-scale population evacuation from urban areas may occur during disasters such as earth quakes, volcano eruptions, militarized conflicts, environmental disasters and more. Efficient and safe population evacuation is of great importance as it can save lives and reduce human suffering. The current study demonstrates an Agent-Based Simulation tool which may be used to support operational planning for population evacuation from threatened urban areas. The simulation models households as agents, each acting in accordance to a designated decision function, which renders the probability of evacuation as a function of the socioeconomic and demographic characteristics of the agents and the behavior of neighboring ones. Upon evacuation decision, agents embark on their way to their preassigned destinations, while their optimal route is calculated and updated periodically, based on road information (taken from Open-Street Map), accumulative traffic congestion, and simulated road conditions. The simulation calculates and records the location of all agents and enables the user to identify and analyze different evacuation scenarios, compare evacuation sequences, map and identify road bottlenecks, etc. Integrating such a simulation into the planning process—both at the municipal and the national levels—can significantly enhance authorities’ processes of preparing evacuation plans, including investing resources for developing safe evacuation destinations and educating the population for the unfolding of future emergencies. The main contribution of the current work is the ability to efficiently calculate optimal routes for millions of agents. To demonstrate this we applied the simulation on Kyiv (Ukraine), where a large number of its 3 million citizens have fled the city during Russia’s invasion on February 2022.
We want to understand in which circumstances identity fusion occurs. We propose a model in which individual needs and interactions between agents and their social environment come together. We argue the personal identity of an agent will fuse with a group when it has a high need for significance and he is member of a group providing a means to gain significance. Agents cannot join all groups to meet their needs, as agents need to have a social connection with the group and need to be accepted within the group. The model allows for multiple scenarios to occur. Agents with a need for significance not necessarily become fused and will find alternative ways to satisfy their need.
Humans have an intrinsic need for friendship, especially in adolescence when entering a new social environment where they do not know anybody. The question as to how friendships form is frequently asked. In research, three important factors have been identified in the formation of friendship: extraversion, resemblance and social status. To our best knowledge, a missing aspect in current research on friendship formation is the concept of “social battery”. The social battery comprehends an individuals’ energy level to engage in social contact. When the social battery is exhausted, it can prevent an individual from social contact, and consequently from making new friends. The recharging and exhaustion of the social battery heavily depends on the person’s extraversion level. In this paper, we develop an agent-based model “the Friendship Field” that simulates real-life dyadic friendship formation where the individuals’ interactions are motivated by their social battery. With this model, we investigate emergent patterns regarding extraversion, resemblance and status. The model reproduces a pattern of the mere-exposure-effect, an existing theory on friendship formation. Moreover, it proposes a new factor for friendship formation in social sciences: the social battery.
This paper investigates the possibility of simulating bounded rationality effects in an agent’s decision-making scheme by limiting its capability of perceiving information and utilising a decision-making framework of Triandis’ Theory of Interpersonal Behaviour. Based on previous work on an agent-based platform, BedDeM, we propose how to capture the effects of sequential, emotional, habitual and multi-criteria decision-making. The Perception component in the agent is further extended to take into account confirmation bias and the bandwagon effect. We demonstrate the functionality of this model in the context of purchasing vehicles in Switzerland’s households.
The main goal of this research is to study the dynamics of the opinion among teenagers by reconstructing the processes of influence that take place during their interactions, raising further questions about the ways and reasons why individuals get in touch with others. The integration between Agent-Based Modelling and sociometry allowed the conceptualization of the phenomenon as a diffusion study, considered as the outcome of the imitative process triggered by any compliance motives especially in view of the sociological tradition. In particular, the concepts of social influence and homophily can be traced back to a dual mechanism able of explaining it: (1) the behavior of peers occupying a relevant position within relational groups (school classes); (2) the interaction favored by specific elements linked to the similarity between individuals. The empirical results obtained from a web survey has been compared with the ones from the simulation model in order to reproduce the above social phenomenon and to confirm the theoretical assumptions behind the model itself.
The transition of heat provision in the urban building stock towards climate neutral sources poses a major challenge to German cities. The underlying actor structure is complex and interlinked. Municipalities set regulatory boundary conditions and decide on infrastructure investments like district heating networks. Necessary investments on the premises of house owners are not only inhibited by unavailability of capital but also by a lack of technical knowledge and ultimately by capacity shortages of installation companies. In the paper, we outline the agent-based model being developed in the course of a new research project aiming to support local heat transitions by socio-technical modelling and simulation. We aim to represent the investment dynamics evolving from interactions of building owners with a broader set of stakeholders, namely energy consultants whose knowledge and thus recommendations shape the set of investment options, and craftspeople such as plumbers whose experience has an impact of building owners’ decisions. We outline the integration of the agent-based model with a model of the local energy system to account for feedbacks between the heating infrastructure and investment decisions of building owners. Furthermore, we discuss our approach to auto-parameterise intervention measures to achieve required rates of investments.
The movement of people in the city varies significantly during the day. However, the availably of open localization data that could be useful in calibration of pedestrian ABM is negligible. The investigation of pedestrian traffic fluctuations could be an important element of city management (e.g. planning public transport, identification of bottlenecks). For that reason, the paper develops the agent-based model of pedestrians’ flows dynamics in the center of one of the largest Polish cities (Poznan). The Google Places traffic data as well as census data and Geographical Information System were used to calibrate the model to generate reliable fluctuations of pedestrian movements. The developed ABM provides several valuable information that stand behind aggregate Google Places popular times rank. Mainly, we estimated the speed and size of pedestrians’ flows together with the inflow and outflow of pedestrians to the city center. We were also able to identify bottlenecks, pedestrians’ waves and areas of high/low density. The model captures and confirms several facts associated with fundamental diagrams of pedestrian flow and it could be used for further experiments regarding urban planning.
Despite being understated, anxiety is a critical factor affecting all levels of society, directly impacting individual decisions and with well-identified ramifications on social play, social constructs, and collective outcomes, as well as being a significant direct social toll tied to yearly trillion-USD social cost. Through a systematic literature review of social simulation research featuring models of anxiety, this paper frames the state of the art on anxiety modelling, and identifies trends and patterns in bibliographic indicators, aspects of anxiety that are modelled, how they are modelled, and their purpose and integration within agent based models. Based on these findings, this paper proposes a way forward as to structure the field as to enable the social simulation community as a whole to cover this critical aspect.
For exploration of future transition paths of the energy system and the complex challenges related to it, modeling components that are either a part of or connected to the energy system is primary. Here, co-simulation approaches facilitate integrated simulation scenarios by coupling simulation models developed in different programming languages, based on different modeling paradigms, and depicting various domains of the energy system (e.g., industry, households, or the electricity grid). However, co-simulation approaches exhibit a range of challenges and are thus under-exploited when investigating socio-technical transitions. We introduce a design and modeling process for an agent-based co-simulation framework, which aims to foster interdisciplinary collaboration considering multiple socio-technical elements of the energy system. This starts with building an information model for simulation planning and collecting inputs and outputs of different models. Finally, a modularization approach defines simulation sub-scenarios to simplify modeling interdependencies. Additionally, we present two exemplary scenarios: (i) the impact of households‘ energy-related behavior on power grid stability and (ii) the co-evolutionary supply and demand dynamics of energy storage technologies in the industrial sector.
Trust is crucial in economic complex adaptive systems, where agents frequently change the other side of their interactions, which often leads to changes in the system's structure. In such a system, agents who seek as much as possible to build lasting trust relationships for long-term confident interactions with their counterparts decide whom to interact with based on their level of trust in existing partners. A trust crisis refers to the time when the level of trust between agents drops so much that there is no incentive to interact, a situation that ultimately leads to the collapse of the system. This paper presents an agent-based model of the interbank market and evaluates the effects of using a voting-based consensus mechanism embedded in a blockchain-based loan system on maintaining trust between agents and system stability. In this paper, we rely on the fact that blockchain as a distributed system only manages the existing trust and does not create it on its own. Furthermore, this study uses actual blockchain technology in its simulation rather than simply presenting an abstraction.
This paper outlines a computational, cognitive model representing how humans may use epistemic vigilance to evaluate socially-provided information in a way that reacts flexibly to differences in the reliability of content versus source vigilance strategies. Furthermore, the model explores how the system reacts in situations where the utility of the information provided is either unrelated to its accuracy or, even, is inversely proportional to it. We find that even a simple model is able to react flexibly to variation in these parameters, providing a basis for further exploration of the phenomenon.
The concept of System of Systems (SoS) is important to realize a society that creates sustainable value through the coordination and cooperation of systems. However, for SoS consisting of subsystems at different spatio-temporal levels, conventional modeling methods are applied independently at each level, making it impossible to conduct both macro and micro evaluations at the same time. In this paper, we propose a multi-scale modeling method that enables modeling of each system component at different spatio-temporal levels. The proposed method is applied to a local city under COVID-19, and a comprehensive analysis of the target system is conducted by modeling and integrating decision-makers at different levels: citizens, organizations, and municipality.
We explore microsimulation design options as a source of divergence in total population when using demographic statistics from the United Nations to model population dynamics in three countries between 1950 and 2100. We compare 176 unique model designs, which toggle options such as the time step, the initial sample size of agents, variance reduction, ordering of demographic events, and adjustments to risk assignment as appropriate to each statistic. Results indicate that small population samples and 1-year time steps can produce particularly high divergence from UN targets, even when other options known to reduce divergence are implemented. Small sample 1-year models with low divergence are possible, but the specific combinations of options interact with a country’s population dynamics in unpredictable ways, which prevents the design from being used in other country contexts. These findings are important for balancing efficiency, accuracy, realism, and generalizability in demographic microsimulation design.
This paper introduces a model of multi-unit organizations with either static structures, i.e., they are designed top-down following classical approaches to organizational design, or dynamic structures, i.e., the structures emerge over time from micro-level decisions. In the latter case, the units are capable of learning about the technical interdependencies of the task they face, and they use their knowledge by adapting the task allocation from time to time. In both static and dynamic organizations, searching for actions to increase the performance can either be carried out individually or collaboratively. The results indicate that (i) collaborative search processes can help overcome the adverse effects of inefficient task allocations as long as there is an internal fit with other organizational design elements, and (ii) for dynamic organizations, the emergent task allocation does not necessarily mirror the technical interdependencies of the task the organizations face, even though the same (or even higher) performances are achieved.
One of the core assumptions made when building agent-based simulation models is how the agents decide or reason about the action to take next. The mode of reasoning is usually the same for all agents and over time within the simulation run. However, is this adequate? There exist several frameworks that describe multi-mode reasoning, however how do we know what we need? To engage with this core question, we reflect on this modelling process, by using CAFCA—one of these multi-mode frameworks—and reflect on the reasoning dimension in a social dilemma decision situation. More specifically, using existing qualitative inquiry on group dynamics in a common pool resource dilemma—not designed to elicit different types of reasoning—we introduce our hunt for reasoning hints and reflect on what insights/data we would need to make an informed decision about the reasoning(s) in our modelling and how to obtain this.
This paper is an attempt to study a well known (probably little studied) phenomenon in academia: citation cartels. This is the tacit or explicit agreement among authors to cite each other more often than they would do in a more "sincere" approach to science. It can be intended as collusion and it can distort scientific progress in affecting a scholar's attention. The phenomenon has been around for decades and it does not seem to spare any discipline. By starting from outlining the characteristics of a "cartel," this study then builds an agent-based model in an attempt to define the extent to which colluding behavior affects progress in a given discipline by operating on citation counts. Data is still preliminary although enough to conclude that cartels promote lax scientific practices.
The emergence of new organizational forms—such as virtual teams—has brought forward some challenges for teams. One of the most relevant challenges is coordinating the decisions of team members who work from different places. Intuition suggests that task performance should improve if the team members’ decisions are coordinated. However, previous research suggests that the effect of coordination on task performance is ambiguous. Specifically, the effect of coordination on task performance depends on aspects such as the team members’ learning and the changes in team composition over time. This paper aims to understand how these two factors moderate the relationship between coordination and task performance. We implement an agent-based modeling approach based on the NK framework to fulfill our research objective. Our results suggest that both factors have moderating effects. Specifically, we find that excessive individual learning harms the task performance of fully autonomous teams but is less detrimental for teams that coordinate their decisions. In addition, we find that teams that coordinate their decisions benefit from changing their composition in the short term, but fully autonomous teams do not. In conclusion, teams that coordinate their decisions benefit more from individual learning and dynamic composition than teams that do not coordinate. Nevertheless, we should note that the existence of moderating effects does not imply that coordination improves task performance. Whether coordination improves task performance depends on the interdependencies between the team members’ decisions.
In social modeling, a computational environment runs a model that represents the world. The states the model explores (its behavioral attractor) are typically fewer than its description suggests. The mapping between model and attractor depends not only on its parameters (exploring variants of the world) and its conventions (imposed by the computing environment), but also its mechanisms (components of the model representing selected dimensions of the world). We illustrate the impact of different mechanisms on the attractor. In our case, in general, the more mechanisms one implements, the smaller the attractor (“the more you model, the less you see”), but with unexpected twists.
Laboratory experiments are among the most frequently used methods in management accounting research because they offer high internal validity, enabling the examination of causal relationships. However, experiments often struggle with providing support for a specific proposed causal mechanism, given the abundance of psychological and behavioral theories that predict similar outcomes. In this paper, we argue that agent-based modeling is well suited to complement experiments because agent-based modeling is a powerful method to increase confidence in the proposed causal mechanism. As a showcase project, we conduct an experiment to explain antecedents of honest reporting behavior in a participative budgeting setting and propose that a social norm of honesty is the underlying causal mechanism. Next, we adapt an agent-based model to our participative budgeting setting and create two submodels incorporating alternative causal mechanisms. Finally, we assess the capability of the two submodels to reproduce the experiment’s results to evaluate whether the observed behavior in the experiment can be better explained with the causal mechanism representing social norm theory.
Agri-environment schemes (AES) are government-funded voluntary programs that incentivise farmers and land managers for environmental friendly farming practices. Understanding farmers’ decision-making process and its impact on AES adoption can aid policy makers in designing AES schemes that meet adoption goals and environmental targets. Farmers’ decision-making is complex and involves a range of social, behavioural, economic and ecological factors. In this paper, we present a spatially explicit agent-based model (ABM)—BESTMAP-ABM-UK that simulates farmers’ decision-making process, inclusive of farmers’ social, behavioural and economic factors, on adopting buffer strips, cover crops, grassland management and arable land conversion to grassland schemes in the UK. The model produces farmers’ AES adoption under varied AES scheme designs in term of the contract length, the offered payment level, the bureaucracy level and the required minimal area. We apply the Morris screening method to analyse the importance of the model parameters in a status quo scenario, in which current UK AES designs are used. The results show that the average accepted payments of farmers for buffer strips and grassland management and farmers’ intrinsic openness to buffer strips have the most significant impact on the farm adoption rate in the model.
