Programme

Topics. In this tutorial, we would like to explain the needs of synthesized population for whole population in a country. When we try to develop a social simulation tool for a real community, we need detail compositions of the real population in that community. In these years, we have synthesized whole Japanese populations that includes compositions of each household. We have synthesized the Japanese population based on the national census in 2000, 2005, 2010 and 2015. That is, we have released four synthesized populations according to the national census. Since we have developed the synthesized method employing a simulated annealing method, the synthesized population depends on some random numbers. Therefore, we have prepared 100 sets of synthesized populations for each census. Researchers can employ all the 100 sets when they employ their simulation tool with the synthesized population. In this tutorial, we would like to explain how to synthesize the populations using the national census, and the rules for distributing the synthesized populations for researchers. We also explain how we can utilize the population in simulating in some area of real-scale social simulations. We would like to call for researchers who are interested in synthesizing the nation-wide populations for real-scale social simulations for their countries.

For info, contact the chair here

Topics. This tutorial introduces the recently published MBSSM (Mechanism-Based Social Systems Modelling) software architecture. The MBSSM architecture is designed for expressing mechanisms of social theories with individual behaviour components in a unified way and implementing these mechanisms in an agent-based simulation model. The architecture is based on a middle-range theory approach most recently expounded by analytical sociology and is designed in the object-oriented programming paradigm with Unified Modelling Language diagrams.
The tutorial presents two examples of using the architecture for modelling individual behaviour mechanisms that give rise to the dynamics of population-level alcohol use: a single-theory model of norm theory and a multi-theory model that combines norm theory with role theory. The tutorial also briefly discusses the research front of “structural calibration” (described variously as “theory discovery”, “model discovery”, “model crunching”, and “inverse generative social science”). The attendees should recognize a fundamental enabling role of the architecture within a wider simulation model-based framework of abductive reasoning in which families of theories are tested for their ability to explain concrete social phenomena.

The detailed programme of the tutorial is here.

For info, contact the chair here

For info, contact the chair here

The detailed programme of the tutorial is here.

For info, contact the chair here

Topics. Agent-based simulation can be related to qualitative as well as quantitative data. For example, qualitative input might be used to inform the micro-level specification of agent behaviour in simulations that are then run and compared to aggregate quantitative data. However using qualitative data can seem daunting, partly because there are no established methods for doing this. The goal of this SSW2020 session is to discuss methods for integrating qualitative and quantitative data in agent-based models, with reference to worked examples.

For info, contact the chair here

Programme Part I

10:00-10:05 Welcome
10:05-10:25
Bruce Edmonds (Manchester Metropolitan University, UK) – “An introduction to using qualitative data for informing simulation design”
10:25-10:55
Keynote speaker: Juliette Rouchier (Université Paris-Dauphine, France) – “Quali and quanti research walking hand in hand”
10:55-11:00 Break

Note: Besides registering for the Social Simulation Week 2020, please register for your participation in this workshop by September 10th using this link: https://forms.gle/1MUTSYGA4BxYD3DX7.

The workshop will feature four short presentations followed by an open discussion. Each of the presentation brings in a different approach (spatial optimisation, adaptive behaviour, innovation diffusion as well as an experimental protocol) and finds itself at different places of two analytical postures:

Relation between data and theory. Two of them are more data-grounded and two more theory-gronded.
Agrifood transitions. Each one deals with a different aspect, i.e. the territory and its farmers, consumers as individuals and their social dynamics, interactions within a specific industry.

Programme

11:15. Welcome organic coffee and healthy, local, and responsibly sourced snacks (bring your own!)
11:30. Introduction to the workshop
11:35. Presentations:

GeoPAT: a spatially explicit tool for prospecting reallocation of cultural patches – Nicolas Dumoulin (Laboratoire d’Ingénierie pour les Systèmes Complexes, INRAE, France)

You are what you eat! Using Experiments to Study (Online) Social Influence Effects on Food Choices – Adrian Lueders (Laboratoire de Psychologie Sociale et Cognitive / CNRS) Université Clermont Auvergne, France)

Modeling Extended Agro-Food Supply Chain: Pathways to Sustainability through Consumer Behavioral Change – Firouzeh Taghikhah (Center on Persuasive Systems for Wise Adaptive Living, Faculty of Engineering and Information Technology, University of Technology Sydney, Australia)

An agent-based model of (food) consumption: Accounting for the Intention-BehaviourGap on three dimensions of characteristics with limited knowledge
– Pedro López Merino (LAMSADE, Université Paris-Dauphine, France)

For info, contact the chair here.

For info, contact the chair here

Programme Part II

11:00-11:20
Stephanie Dornschneider (University College Dublin, Ireland) “Building bridges: Connecting ethnography to agent-based modelling”
11:20-11:40
Martin Neumann (Johannes Gutenberg-Universität Mainz, Germany) – “Hermeneutic interpretation of simulation results”
11:40-11:45 Break
11:45-12:05
Sukaina Bharwani (Oxford Centre, UK) – “Adding value to social simulation models using qualitative evidence”

Note: Besides registering for the Social Simulation Week 2020, please register for your participation in this workshop by September 10th using this link: https://forms.gle/1MUTSYGA4BxYD3DX7.

12:20. Discussion chaired by Juliette Rouchier and Sylvie Huet

If you’d like to participate, please register to the SSW 2020 first and then fill in the following form (to be closed at 12pm the day before the Workshop):
https://forms.office.com/Pages/ResponsePage.

For info, contact the chair here.

12:05-12:25
Patrycja Antosz (University of Groningen, The Netherlands) – “Juggling quant, qual and simulation: lessons learned in the SMARTEES project” [Co-author: Wander Jager] 12:25-12:45
Fishbowl discussion, moderator: Melania Borit (UiT The Arctic University of Norway, Norway)
12:45-1:00
Plan future events, moderators: Melania Borit and Bruce Edmonds

Note: Besides registering for the Social Simulation Week 2020, please register for your participation in this workshop by September 10th using this link: https://forms.gle/1MUTSYGA4BxYD3DX7.

For info, contact the chair here

Topics. Bridging micro- and macro-scales is pivotal to both social network research and agent-based modelling (ABMs). Exchanges between SNA and ABM have recently increased. ABMs can explain how network dynamics and macro-level outcomes can be linked through micro-level mechanisms. The network component of an ABM can be empirically calibrated, which allows ABM modellers to move beyond the use of abstract networks in early ABMs. Recent developments in statistical models of repeated network observations have brought up discussions on the use of simulations in social network research. Moreover, ABM can be used as tools to increase the reliability of statistical analysis of network data. This session invites papers that rely on ABMs relating macro-level outcomes with social network dynamics in e.g., opinion polarization, social inequality, social conflicts, or economic collaboration. Particularly — but not exclusively — welcome are contributions bridging theoretical ABM, empirical data and statistical models of network generating processes (e.g., ERGM, SAOM).

For info, contact the chair here

Programme

2PM
Micro-macro modelling with network models designed for empirical inference – Christian Steglich (University of Groningen and Linköping University)

2:45PM
From empirical to hypothetical: A model for political opinions calibrated by SAOM – Kieran Mepham (ETH Zürich)
Empirical network analysts and agent-based modellers alike have recently called for advances in bridging of these two fields. In our current work, we aim to take steps towards this by using a SAOM as an empirical calibration for an agent-based model of political attitude and social clustering. We model multiple political attitudes, treating them as a valenced two-mode network of individuals and topics. Then, we model the coevolution of this network with the friendship network between individuals. This model is based on a previous project applying SAOMs, in which we collected and analyzed data in the scope of the Swiss StudentLife Study, estimating the strength of micro processes of selection and influence on individuals’ choices, and how these ultimately related to meso-level structural outcomes. We use this model as the basis for forward simulation. As a consequence of the SAOM framework applied, our agent-based model assumes one-to-one influence with randomly distributed sequential updates. Selection of friends is modelled as a stochastic process, where each additional shared opinion contributes equally to a linear predictor of a logistic function, weighting the relative chances of adjusting one (potential) outgoing tie over another. Similarly, the choice to adopt a political opinion is modelled through a linear predictor which increases equally for each additional friend who already holds this opinion. Additional parameters model unobserved ties and contextual factors which may drive convergence and divergence between individuals, endogeneity in the friendship network, and friendship homophily on demographic characteristics. We vary the focal selection and influence parameters and starting network configuration to examine their consequences, and trial various metrics to see how these changes affect the consequences of micro-processes in the long run. We also offer some points on conceptual differences between this approach and more traditional agent-based models.

3:10PM
Bad barrels spoiling good apples in social dilemmas: How social network information can brighten a dark side of meritocratic matching – Carlos de Matos Fernandes (University of Groningen)
Ideally, past achievements and performances, conceptualized as individual prior merit, effectively signal individuals’ abilities. Individuals gain access to better groups based on merit. But this is particularly information-heavy in situations when there is a large pool of individuals. In most cases, we must rely on incomplete information such as group merit. Where and with whom you worked in the past may be indicative of your individual merit and has less severe information assumptions. But relying on group merit poses a problem on its own: the bad barrels problem. Groups in which defection prevails are labelled as ‘bad barrels’. A ‘good apple’ is an actor who is prosocial and wants to cooperate. Still, we infer from reciprocity and learning principles that actors, even prosocials, are more likely to defect when others do so as well (Axelrod, 1984; Macy & Flache, 2002). Such actors are then ‘spoiled’. The problem arises in situations where groups decide to select new members partially based on information about candidates’ prior group performance. ‘Spoiled’ actors may get stuck in uncooperative groups. In this paper, we explore by means of agent-based modelling under which conditions local information via network relations, cutting across group boundaries, can help prosocial actors escaping uncooperative groups. We assume that in past dyadic interactions agents learn about others’ types (Buskens & Raub, 2013). We want to know under which structural settings such networks may buffer the bad barrels problem during meritocratic matching. We analyze a stochastic learning model with adaptive thresholds (Macy, 1991). Generally, if cooperation (defection) generates a positive individual outcome, thresholds decline (increase), making cooperation more (less) likely. Agents are classified as either prosocial or proself. Prosocials have initially lower thresholds than proselfs. The basics of matching are as follows. All agents are randomly grouped, playing iterated n-person prisoners dilemmas (INPDs) for x rounds. Unsatisfied agents have the opportunity to leave a group after x rounds. Ungrouped agents are then sorted based on merit (depending on information available) and matched with others with similar revealed group performance, followed by playing INPDs in new groups for another x rounds. Agents are embedded in a network, generated via a spatial random graph algorithm (Wong, Pattison, & Robbins, 2006). Agents play 2-person PDs with one of their network partners, eventually generating a social reputation. Next to a random network, we are particularly interested in the effects of network clustering in combination with prosociality homophily because this affects the degree to which good apples can effectively signal others about their prosociality. We apply the method of decreasing abstraction – from complete to incomplete information – to vary the information agents have during matching. We implement 6 matching rules, starting from 1 (complete information) to 6 (incomplete information and local social network information). Proposition 1. Bad barrels spoil good apples in situations when meritocratic matching to join potential new groups is based on actors’ prior group performance. Proposition 2. The possibility to signal one’s prosociality in dyadic interactions weakens the bad barrels effect by improving prosocials’ chances to escape bad barrels. Proposition 3. Homophily buffers the bad barrels effect due to network clustering and by promoting within-cluster interactions between prosocials, thereby facilitating the identification of prosocial types during matching.

3:35PM
Cultural and Opinion dynamics in small-world “social” networks – Yunsub Lee (Cornell University)

Agent-based models of cultural and opinion dynamics have explored how, in social networks, a simple process of interpersonal influence between actors leads to the emergence of cultural consensus or political polarization. Despite their explanatory power on the micro-macro linkage, previous models reflect only the topological features of social networks (e.g., a network structure is random or small-world), missing its relational features that are characterized by various social entities, such as norms and institutions. In this study, we suggest two formal parameters, which are applicable to most previous models, reflecting two relational features of social networks: the strength of norm-controlled ties (i.e., ties in dense connections survive longer than in sparse connections due to the norms enforced) and induced homophily (i.e., institutions, such as job, school, and neighborhood, yields network segregation by gender, race, and class). We apply both parameters to a previous model (Flache and Macy 2011a), maintaining all other simulation conditions the same. Unlike the previous model that reveals more polarization by more long-range ties in a small-world network, our updated model produces more complex and unexpected outcomes—random emergence of extreme polarization or consensus. However, when the parameters are applied separately, the outcomes show no difference from the previous one.

4:00PM Closing

Topics. The event proposes to present the most recent work on simulations in Economics of some LatinAmerican research groups working in the area of Complex Systems, Agent-based Models, Nonlinear Dynamics and related topics. The collection of papers to be presented include some theoretical models that require the use of simulations, as well as empirical exercises applying simulation models to data. The individuals and research groups proposing the virtual workshop is an active group in LatinAmerica that works on the area of Complex Systems and particularly in Agent Based Models, with emphasis on issues related to emerging markets and the links between the economy, production systems and underlying socio-ecological systems.

Presentations

Martha Alatriste-Contreras (Universidad Nacional Autónoma de México, México): “The impact of sectoral shocks in the North America Production Network“. [co-authors: Alatriste-Contreras, M. & Puchet, M]

Abstract: In the context of the new trade agreement between the countries of North America (TMEC acronym in Spanish), it is important to evaluate the posible impact of sectoral shocks in the North America Production Network. We use input-output data for Canada, Mexico, the USA, and the region, a network diffusion model, and computer simulations to evaluate the effect of specific sectoral shocks in the countries’ and region’s economy. The diffusion model we use assumes that a sectoral shock changes the input-output connections between sectors, thus changes the way sectors produce. The results of the simulations shed light on the posible impacts of the new trade agreement and provide useful information to design an industrial policy focused on the development of the production network. In particular, we focus on recommendations for the Mexican economy.

Marcelo Álvez (Universidad de la República & Central Bank of Uruguay, Uruguay): “Progressive income tax and its emerging growth effects: a complex systems approach” [Co-authors: Alvarez, E., Álvez, M. & Brida, J.G.]

Abstract: The State as a distorting agent of the markets or the State as a corrector in the face of market failures, these are two characterizations that allow sizing the diversity of positions regarding the role of this particular agent of the economy. In this work, an agent-based stock-flow consistent model (AB-SFC) is applied to analyze the differences in the economies when establishing different types of taxes on personal income, proportional and progressive. Different combinations of threshold and rate are tested. There are no significant differences in economic performance in the presence of one tax scheme or the other. In the scenarios with higher rates, a slower recovery of the economy is noticed after a period of stagnation, but the difference is not significant at 90% significance. This design, which only distinguishes two sections of income, is not able to reduce the inequality generated throughout the income distribution. The tax design seems to offset the inequality in the lower section of income distribution through tax exemption for low-income households, but not the one generated in the section of higher income. An additional policy is necessary to offset the differences generated in the range of higher-income individuals. In this exercise, there is no evidence of a deterioration of economic growth in the presence of a progressive income tax, instead of a proportional one.

Emiliano Alvarez (Universidad de la República, Uruguay): “Agent Based Models and Simulation in Social Sciences: A bibliometric review” [co-authors: Alvarez, E. & Brida, J.G.]

Abstract: Since the first agent-based models (ABM), the scientific community has been interested in making not only the results of computational models understandable but also the modeling description, to facilitate their replication. The form that has been adopted to a greater extent has been the ODD (Overview, Design concepts, and Details) protocol, which provides a generic structure for its documentation. This protocol provides a way to clearly explain the procedures and interactions of the complex systems to be analyzed, with applications that have spread across
different disciplines. This work will show a bibliometric review of the articles that emerged from the first publication of this protocol in 2006, analyzing the development that ABMs have had in the social sciences. A description will be made of the lines of research with the greatest activity and the links between them will be analyzed; while summarizing the countries, universities, and journals with the highest contributions.

Nicolás Garrido (Universidad Diego Portales, Chile): “Crowding and price dynamics in tourism destination choice” [co-authors: Alvarez, E., Brida, J.G. & Garrido, N.]

Abstract: This paper analyzes how the preferences for crowding in destinations by tourists interact along the time with the pricing strategies of the resorts to determines the number of tourist visiting a tourist attraction. Destinations are experience goods and the stakeholders of the destinations use multiple signals to reduce the uncertainty of the consumers before their choice. Taking into consideration companies that seek to increase their profits and customers with budget restrictions and with both individual and social preferences, this work analyzes price dynamics and conditions of market competition, under an Agent-Based Modelling (ABM) setting. An emerging result of this process is the formation of companies with greater market power and high customer differentiation, although less than in the case without budget restrictions. At the same time, adjustment speed of the companies have non-linear effects on prices, the benefits of the resorts and individuals’ utility. Moreover, price dynamics is highly sensitive to the set of information that agents in the market have.

Daniel Heymann (Universidad de Buenos Aires, Argentina): “Behavioral heuristics and market patterns in a Bertrand–Edgeworth game” [co-authors: Heymann, D., Kawamura, E., Perazzo, R. & Zimmermann, M.G.]

Abstract: This paper studies Bertrand price-setting behavior when firms face capacity constraints (Bertrand–Edgeworth game). This game is known to lack equilibria in pure strategies, while the mixed-strategy equilibria are hard to characterize. We explore families of heuristic rules for individual price-setting behavior and the resulting market patterns, through simulations of agent-based models and laboratory experiments. Overall, the individual pricing strategies observed experimentally can be represented approximately by a sales-based simple rule. In the experiments, average market prices tend to converge from above and approach a state resembling a steady state, with slow aggregate price variations and low price dispersion around an average near the competitive level. However, that configuration can be disturbed occasionally by excursions triggered by discrete price raises of some agents. Salient features of experimental results can be described by simulations where agents use sales-based heuristics with parameters calibrated from the experiments. The results obtained here suggest the existence of useful complementarities between analytical, experimental and agent-based simulation approaches.

For info, contact the chair here

Topics. This event aims to provide some ‘behind-the-scene’ information on JASSS and allows interested authors, referees and readers to discuss the journal development with the editor.

For info, contact the chair here

Topics. Following the editorial “Computational Models That Matter During a Global Pandemic Outbreak: A Call to Action” in JASSS 23(2), there have been ample reactions on the RofASSS site (see https://rofasss.org/tag/jasss-covid19-thread/) indicating a large interest within the social simulation community on the modelling of global pandemics and COVID-19 in particular. Ranging from the value of modelling in general or for policy makers and making, criteria for transparent modelling, development of simulation models and tools, experiences and ideas for fast model adaptation and development in a crisis situation, KISS versus KIDS, etc. We assume this broad and varied interest is a solid base for a Social Simulation Week workshop to extend the current asynchronous discussion with synchronous discussion and possible the start of integrative cooperation as well as deep epistemological debate. We aim to invite some guests from outside the social simulation community who have worked with individual-based epidemiological modelling..

Programme

Session 1
10.00-10.55AM Social simulation and epidimological models in the times of COVID-19

Christopher Watts (independent): An agent-based modeller looks at an epidemiologist’s model
Nick Malleson (Geography, University of Leeds): A microsimulation model of COVID spread in the UK
Fredrik Liljeros (Sociology, Stockholm University): The importance of contact patterns to counteract blackbox models
Jennifer Badham (Sociology, Durham University): Justified stories for local COVID-19 planning
Imran Mahmood (Computer Science, Brunel University): Comparing ABM approach with conventional approaches in Disease Modeling

Session 2
11.00-11.35AM ABM models in the times of COVID-19

Nick Gotts (Geography, University of Leeds ): Agent-Based Modelling of Covid-19 Transmission in Hospital Settings: Plans in the SAFER Project
Alexis Drogoul (Institute of Research for Development, Marseille): COMOKIT
Frank Dignum (Computer Science, Umeå University): The ASSOCC model

Session 3
11.45-12.15AM Policy and ABM in the times of COVID-19

Jason Thompson (Melbourne School of Design): Bringing computational social science into mainstream policy making for COVID
Pietro Terna (University of Torino, Italy, retired, and Collegio Carlo Alberto, Italy): How can ABM models become part of the policy-making process in times of emergencies
Andreas Tolk (The Covid-19 Healthcare Coalition M&S Group): The Artificial University (TAU)

Session 4
12.25-12.55AM Crisis and ABM

Juliette Rouchier (LAMSADE, Paris): Information and modelling during crisis
F. LeRon Shults (NORCE Center for Modeling Social Systems, Norway): Emotional Contagion: The Spread of Misinformation, Stigma, and Fear during a Pandemic

For info, contact the chair here

Programme

2PM
Network mechanisms and network models – Christoph Stadtfeld (ETH Zürich)
Societies can be understood as complex systems of interdependent social actors that are tied to one another in various ways. While the individual actors may have limited understanding of the system as a whole and mostly respond to their local environments (the micro-level), their interdependent actions shape macro-level features that in turn affect the opportunities of actors to act and interact. The micro-macro motif is a central concept in analytical sociology (Hedström &Bearman 2009; Coleman 1990). Social networks are a formal framework to describe, analyze, model and theorize about such interdependent social systems. In particular, social network concepts are suitable to describe both micro-level processes and their macro-level properties (Stadtfeld 2018). This talk introduces strategies for the analytical investigation of social mechanisms in social networks. The goal is to explain structural properties and change in social networks. It first defines the concept of network mechanisms (e.g., reciprocity, transitivity, popularity) and compares and contrasts those with the definition of social mechanisms by Hedström & Swedberg (1998). It then focuses on how those network mechanisms can be tested with statistical network models in a variety of empirical settings and with different types of network data. It further discusses how some of these statistical models may be applied as simulation models to understand the role of network mechanisms in emerging macro-level network properties.

2:45PM
Social Network Dynamics and Sustained Grading Discrimination: An Agent-Based Model – Károly Takács (Linköping University)
School grades are important determinants of educational attainment. Several empirical studies found evidence for a gap in grades for different groups of students after controlling for blind and unbiased test scores. Girls, for instance, were found to receive better non-blind assessments than boys and ethnic differences in grading were observed in several countries. This study tries to explore mechanisms that could explain this persistent difference. A possible mechanism has been described as “Acting White”: high performing minority students are pulled back and are threatened by exclusion from friendship due to a pressure arising from anti-achievement norms and oppositional culture. As a different mechanism, their social network ties could create a filter bubble and prevent minority students from improving their situation. Endogenous network dynamics may amplify the negative effects of biased grading as selection into homophilous friendship groups and peer influence on efforts might self-reinforce the grading gap. Moreover, teachers might simply discriminate based on their previous experiences. Even if direct experiences do not exist, differential treatment could be driven by status generalization processes such that members of groups with lower status value in society are judged by stricter standards than others. These mechanisms are hard to disentangle in empirical studies. Using agent-based simulation, we can examine the impact of various mechanisms on the persistence of grading bias. Moreover, we investigate if homophilous friend selection and peer influence amplify the grading bias over time. Our model represents one grade of a school with “red” and “blue” students. Students are arranged in a friendship network and attributed with stable abilities that follow a normal (or empirically informed) distribution. Students receive grades based on their abilities, effort, and an innate stereotype of the teacher. Students update their academic effort based on their direct experience of how their effort was translated into grades. If they feel insufficiently rewarded, they lose motivation and study less. Moreover, their effort levels are influenced by their friends. In every round that could be considered as one grading period, they also update their friendship ties with a certain rate. Structural updates follow mechanisms such as reciprocity, preferential attachment, and transitivity; but ties are also selected based on similarity concerning group membership and academic achievement. From an initial relatively fair distribution of abilities, we do not see the emergence of large within group differences. Our preliminary results do not indicate that clear-cut oppositional cultures could evolve in the lack of statistical differences between the groups and given equal opportunities. As we hardwire in individual development, it is not surprising to see average effort and grades increasing. This tendency towards high effort seems to be strengthened in case of homophilous selection. Low density and tie decay work against overall positive outcomes as they cause segments of low achievers from the same group to emerge. We conclude that just as for opinion dynamics models, it is difficult to explain the emergence of polarized outcomes such as oppositional cultures and persistent grading differences from equal abilities and opportunities.

3:10PM
Homophily in Social Networks and the Propagation of False Information – Jonas Stein (University of Groningen)
Past studies using observational data have suggested that echo chambers (homophilous segments of online networks) boost the spread of false over true information. However, from a theoretical standpoint, one would not necessarily assume that only false information would spread widely in echo chambers. In fact, stronger social influence and confirmation bias should help both true and false messages to perform better in an ideologically aligned homogenous network. In order to explain why past empirical evidence stresses the significance of echo chambers specifically regarding false information, we propose a ‘structural effect’ as a new theoretical argument: Whereas the diffusion of true messages is facilitated by higher general credibility in any network, false messages will only surpass a percolation threshold and diffuse widely if enough susceptible individuals are clustered into homogenous network segments. We illustrate our argument with ABMs in which the diffusion of true and false information for different network types is examined. An experimental online social network study linking insights from the ABMs with empirical data is reported as work in progress.

3:35PM
How extraversion structures friendship networks: An agent-based modeling approach – Alec McGail (Cornell University)
Extraversion is here defined as the amount of interaction a person desires, along two dimensions: time spent with others, and total number of contacts. Through simulation, I investigate how differing preference distributions in the population affect global structural properties, such as degree distribution and the global equilibrium satisfaction people have with their social lives. For example, I observe that lower variation in extraversion in the population leads to less “dissatisfaction” over the population as a whole. To test the robustness of these observations, I perturb the idealistic simulation with various real-world elements, such as social foci, time constraints, differences in tastes and affinities, the use of technology to curb loneliness, differential sociability, disaster or mass-migration, etc. In lieu of discussing the results of all these perturbations in full, I make the code for these available for collaborative input and use, and discuss extension of this code to friendship preferences more generally.

For info, contact the chair here

Topics. The Covid-19 Pandemic has exposed stresses and strains in societies, political systems and sciences. Some governments contend that they entirely “follow the science” (e.g. UK) whilst scientists within and between disciplines show how (seemingly) contradictory recommendations can be. This highlighted interface of policy and science is an opportunity for all kinds of modelling, including agent-based modelling – but not without potential hurdles, bear traps and pitfalls. A ‘call to action’ in JASSS has led to a succession of interesting responses in RofASSS, highlighting the complexity of the endeavour of prediction in complex social systems and using ABM as a suitable method. This session proposes to hold a formal debate on the motion “Agent-based modelling can be used for prediction in complex societal systems”. We will invite a proposer and an opponent for the motion, each of which will speak for 5-10 minutes. We will then seek contributions which, if accepted, will be scheduled in the debate for 3-5 minute talks. At the end of the debate, the proposer and opponent are given a ‘right of reply’, and we will hold a vote on which side won the debate.

Programme

2:00-2:05PM Entry
2:05-2:10OM Introduction to the session by Corinna Elsenbroich
2:10-2:20PM Motion proposed by Edmund Chattoe-Brown
2:20-2:30PM Motion opposed by Bruce Edmonds
2:30-3:15PM Five-minute contributions
– Frank Dignum (Umeå University, on behalf of the ASSOCC project)
– Itzhak Benenson (Tel Aviv University)
– Patrick Steinmann and George van Voorn (Wageningen University & Research)
– André Martins (Universidade de São Paolo)
– Carlos de Matos Fernandes and Marijn Keijzer (University of Groningen)
– Eric Silverman and/or Umberto Gostoli (University of Glasgow)
– Jean-Daniel Kant (LIP6, Sorbonne Université)
– Cesar Garcia-Diaz (Pontifica Universidad Javeriana)
3:15-3:25PM Comfort break
3:25-3:40PM Contributions from the floor
3:40-3:45PM Right of reply by Bruce Edmonds
3:45-3:50PM Right of reply by Edmund Chattoe-Brown
3:50-3:55PM Vote on the motion by all present
3:55-3:00PM Results of the vote and close

For info, contact the chair here.

Topics. ‘Othering’ and polarisation have immediate and potentially severe consequences for politics across Europe – in terms of Populist denigration of sub-groups but also when politics is so divided that each side will not listen to the other (e.g. Brexit in the UK, or on measures to fight Corona, e.g. in Germany). Unfortunately, relevant theory, knowledge and perspectives on these phenomena are splintered across many disciplines, ones that normally do not talk to each other. The Social Identity approach (SIA) refers to the combination of Social Identity Theory (Tajfel & Turner, 1979) and Self-Categorization Theory (Turner et al., 1987; Reicher et al. 2010). SIA proposes that people derive a significant part of their concept of self from the social groups they belong to (Tajfel, 1978; Tajfel & Turner, 1979; Tajfel & Turner 1986; Turner, Hogg, Oakes, Reicher & Wetherell, 1987). SIA proposes that social identification, and the perception of people as fellow group members (or outsiders), is a fundamental basis for collective behaviour. SIA investigates how and when individuals come to feel, think and act as members of a group rather than as individuals. Agent-based modelling (ABM) is a means for bringing the different knowledge and perspectives on these issues together – bringing cognitive and social aspects within a single, coherent framework and sparking interdisciplinary debate. The SIA connects the cognitive to the social outcomes and is amenable to formalisation within ABMs, and is thus one possible means for making simulations. This workshop would present and discuss research on uses of ABM (either existing or prospect) to represent and explore this cluster of phenomena.

See the programme here

For info, contact the chair here

Topics. The coronavirus pandemic has brought agent-based models (ABM) to the attention of public health researchers and policymakers with key government decisions made based on agent-based infectious disease models. However, measures to reduce transmission of the virus have had wide reaching impacts on other aspects of physical and mental health and the economy, highlighting the complex nature of public health issues facing decision makers. In this webinar hosted by the Population Health Agent-based Simulation nEtwork (PHASE), we will discuss potential applications of ABM to address wider public health challenges and highlight key considerations when developing models of public health. Drawing on examples of ABM for adult social care and contact tracing, we will examine issues such as model specification and obtaining suitable data for model calibration and sensitivity analysis. We will also discuss the role of cross-disciplinary partnerships involving health practitioners and decision makers in developing effective and useful models of public health and the ways in which PHASE aims to support these collaborations.

For info, contact the chair here

Programme

10-10.10 Welcome – Dr Alice MacLachlan, University of Glasgow (chair)
10.10-10.25
Prof. Richard Mitchell, University of Glasgow – “Help! Public health needs ABM”.
In this brief talk, I will use the current Covid-19 pandemic to illustrate why public health desperately needs ABMs to help understand and tackle the complex interactions between people and their environment. These interactions are crucial for infectious disease and for the bigger challenge; non-infectious disease. I’ll consider why public health hasn’t used ABM much before now and explore the kinds of questions that could be asked and answered.
10.25-10.55
Dr Jennifer Badham, Visiting Scholar, Queen’s University Belfast – “Help, all my mechanisms are missing”
In this talk, I will argue that a key barrier to widespread adoption of agent-based modelling in public health is that mechanisms are missing from major behaviour theories. A mechanism focus could also help bridge disciplinary gaps.
10-55-11.25
Dr Eric Silverman and Dr Umberto Gostoli, University of Glasgow – “ABM for Social Care Policy”
In this talk, we will present a model of social care provision which we have been developing in the last three years, with the aim to show how ABM can help us to develop models of societies characterized by a complex interaction between demographic, epidemiological and economic factors. We will show how, even in a situation of scarce data, these kinds of models can still be a valuable tool for policy makers to test social and economic policies in order to assess spill-over effects and unintended consequences before these policies are implemented in the real world.
11.25-11.35 Break
11.35- 12.00
Dr Jonatan Almagor and Dr Stefano Picascia, University of Glasgow – “An agent-based model of COVID-19 and the effectiveness of smartphone-based contact tracing”
Using an agent-based model we simulate the transmission of COVID-19 in a population of agents on an urban scale to assess the feasibility of a smartphone-based track-and-trace strategy to mitigate the COVID-19 epidemic.
12.00-12.30
Prof. Laurence Moore, PHASE Network Director, University of Glasgow – “The Population Health Agent-based Simulation nEtwork (PHASE)”
This talk will provide an introduction to the network vision and aims, followed by discussion session about research priorities and network activities.

Programme

2:00PM
Simulating disease spread with statistical network models to design behavioural recommendations – Per Block (Oxford University)
Social distancing and isolation have been widely introduced to counter the COVID-19 pandemic. However, due to adverse social, psychological, and economic consequences of a complete or near-complete lockdown, most restrictions have by now been relaxed and case numbers are increasing again. Adopting a social network approach, we develop and evaluate strategic contact reduction strategies designed to keep the curve flat and aid compliance in a post-lockdown world. We simulate stochastic infection curves incorporating core elements from infection models, ideal-type social network models and statistical relational event models. In particular, we simulate how different choices of face-to-face interaction partners among an individual’s usual social contacts affect infection spread. The underlying social network structure is obtained by simulating ideal-type network models. Thus, simulations have two purposes in the current study: first, they formalise the assumptions about everyday contact structures (simulations to replace unobtainable empirical data); and second, they allow the variation of individual’s interaction strategies (simulations to explore outcomes of potential interventions). Using this approach, we demonstrate that a strategic social network-based reduction of contact strongly enhances the effectiveness of social distancing measures while keeping costs lower.

2:45PM
Do inter-age contacts explain cross-country differences in COVID-19 diffusion? An Agent-Based Model using empirical network data – Luca Sage (Sorbonne University and University of Trento)
Social scientists have argued that side by population age structure, social networks and in particular the strength of intergenerational ties help explaining cross-country differences in the diffusion of covid-19. We argue that due to the high number of confounding factors, it is extremely difficult to test this hypothesis empirically. We use survey data on face-to-face contacts for Great Britain, Italy, and Germany, to reconstruct networks that precisely mirror the empirical inter-age contact patterns as well as the degree distributions. Simulations of the virus diffusion over those networks reveal important cross-country differences in the number of cases across age groups, with Italy being the most hardly hit, followed by Great-Britain and Germany. This holds even when we neutralize the effect of age structure. We ask what characteristics of the empirical networks matter for explaining these cross-country differences, by disentangling the contributions of network density and inter-age contacts patterns. Counterfactual manipulations of networks structures show very little differences due to inter-age contacts patterns. Instead, the high network density of the Italian population causes many cases. Our results suggest, that everything else being constant, interventions to flatten the diffusion curve might be more or less difficult to implement in different countries. Countries with denser social networks might have to enforce more severe limitations on the usual life of persons.

3:10PM
Investigating micro-macro interactions in small-scale fisheries trade networks – Blanca González García-Mon (Stockholm University)
Trade networks are increasingly important for the sustainability of food production systems in a globalized context. Informal trade networks consisting of both economic flows and social relationships are especially important in small-scale fisheries systems. The structure of such trade networks, and the micro-level actions that take place within them (e.g., people fishing, buying, selling), can influence sustainability outcomes. These outcomes will at the same time influence those/peoples’ actions, leading to complex interactions between micro and macro-level processes that are not yet well understood. For example, we hypothesize that trading and fishing decisions will ultimately affect fish availability and the exploitation dynamics of fish populations, which may in turn affect fishing and trading decisions at the micro-level. At the same time, this process is constrained and enabled by social-ecological networks. In this study, we developed a stylized agent-based model based on qualitative and quantitative data from a small-scale fisheries trade system in Mexico, which include the social-ecological network analysis of an empirical fish trade network. We use the model to investigate the following research question: how do different trade network structures influence the sustainability and resilience of fish provision? We specifically look at their influence for three macro-level outcomes: the overexploitation of fish populations, income inequality between different actors, and the availability and variability of fish provision to satisfy different market demands. We analyze how different types of traders in a network adapt to fluctuations of fish supply and demand by changing their target fish species and therefore influencing food provision and the state of fish populations. This model shows how agent-based modeling can be combined with network analysis to better understand the structure and dynamics of trade networks, where different types of fish resources and different types of traders interact influencing macro-level outcomes. More broadly, we use this model as an example to discuss the benefits and challenges of combining agent-based modeling and network analysis for the study of social-ecological systems.

3:45PM
Network Effects on Tax Compliance – Simone Gabbriellini (University of Trento)
Tax evasion threatens government revenues by affecting the quality and sustainability of public and social services for all citizens, including especially those who regularly report their income and those under the poverty line who depend more on these services. Recent reports showed that intentional under-reporting of income is about 18-19% of the total reported income in the US, leading to a loss of about 500 billion USD, while the level of tax evasion in Europe is about 20 percent of GDP, accounting for a potential loss of about 1 trillion each year. In Italy, recent estimates from the Ministry of Economics and Finance said that unpaid taxes were about 270 billion, i.e., about 17% of GDP. To fight this problem, national governments are trying to increase the efficiency and efficacy of audits by collecting more detailed information on taxpayers from a variety of sources, as well as using creative “nudge” experiments to trigger moral instincts and social motivations of taxpayers. However, there is still need for explaining the economic and social determinants of tax compliance and evasion better. This is key also to increase the efficacy and efficiency of tax evasion control policies, which should be higher if options would be more consistent with the understanding of such determinants. Our model assumes a population of N taxpayers with an initial capital, an initial risk propensity, a number of professional contacts to share info with, and the ability to remember (some of) their past, also regarding their neighbors. Remembering facts about neighbors means that the more your neighbors have been sanctioned, the less you will risk, and vice-versa. As it is typical in ABM, our agents are not representative but heterogeneous, adaptive and bounded rational, which means they have, for instance, different endowed beliefs or risk to being caught, and they update these values during each simulation. Also, they differ in income and risk propensity. Our agents are embedded into social networks that allow them to communicate with some of the other agents, i.e. their immediate neighbors. On the one hand, our model emphasizes the role of social imitation, in that the taxpayer’s decision to imitate others (or do the opposite) depend on the recent personal history of the professional contacts. On the other hand, our model assumes that taxpayers build their own weighted audit probability, by looking at their personal history and that of their neighbors simultaneously. The key driver for an agent’s decision to evade is thus a combined outcome of different elements. Networks between taxpayers are our key manipulation in all experiments: we want to estimate the extent of how much topology matters, and what eventually this could imply for policy interventions.

For info, contact the chair here

Topics. Games can aid the development and refinement of agent-based models (ABM) by providing interactive and engaging environments in which social dynamics, perceptions, and behaviours of the players can unfold and be studied. This could then replace or complement lab experiments as well as empirical observations. There are, however, also caveats for the transfer of observations form gameplay to agent-based models since much like agent-based models are an abstraction of reality, so are games. Fine-tuning games to serve as reliable input for ABM development may itself give insights for the fine-tuning of ABM in general and vice versa. Examples of the use of games for ABM development do exist, but the potential is yet to be fully realized. During the workshop, we present a review of established practices and new developments in the field and provide detailed insights from two hands-on methods to integrating games and ABM. All workshop participants are encouraged to engage with the speakers and have the opportunity to articulate their remarks in an interactive session.

Schedule.

Liu Yang (Southeast University, China) will talk about integrating agent-based modelling and serious gaming for planning transport infrastructure and public spaces – 20 min presentation + 10 min discussion.

Break: 5 min.

Timo Szczepanska (UiT – The Arctic University of Norway, Norway) will bring up recent findings of a systematic literature review on established practices and new developments in the field for discussion – 15 min presentation + 15 min discussion.

Break: 5 min

Harko Verhagen (University of Stockholm, Sweden) and Melania Borit (UiT – The Arctic University of Norway, Norway) will initiate an interactive session to explore why we want to mix games and ABM, why we don’t do it, and how we can advance investigating the potential synergies of games and ABM.

Break: 5 min:

Christophe LePage (The French Agricultural Research Centre for International Development- CIRAD-, France) will grant insights to his experiences with applying games in companion modelling – 20 min presentation + 10 min discussion.

Reading material relevant to the session will be circulated in advance. Thus, besides registering for the Social Simulation Week 2020, please register for your participation in this workshop by September 10th using this link: https://forms.gle/hkHMwEWNAvq7M9Yw5.

For info, contact the chair here

Abstract. In many countries, public administrations increasingly use Artificial Intelligence (AI) algorithms to decide on public service provisions among their citizens. Citizen profiles are assessed for their worthiness to receive public services, scoring using value criteria to distinguish between legal /fraudulent recipients, deserving/non-deserving, or needy/non-needy. Although types and degrees of AI implementation vary between countries, delegating decisions about the distribution of scarce resources based on value judgements to machines leads everywhere to important questions of ethics, justice, quality, responsibility, accountability, and transparency of welfare decisions. However, perceptions, attitudes and acceptance of AI use in service provision vary between countries due to different norms and values in-use, different technology status, economic models, civil society sentiments, and legislative, executive and judicial characteristics. Are existing cultural value patterns driving the use of AI, or is AI driving cultural change? What are the impacts of AI for future societies and their value systems? Which policies, behavioural changes and institutional developments are necessary and appropriate to prevent or support certain scenarios? The appraisal of potential social futures is a huge research challenge. This talk presents an approach that uses cultural comparison dimensions to build context-specific ABM simulations for policy advice and testing policy interventions. Simulations work with data at the country level using intelligent agents to model future societies and experiment with AI-in-use for projecting societal techno-futures.

Topics. The generative, or mechanism-based, approach to modeling of social systems uses agent-based models (ABMs) to ‘grow’ the phenomenon under investigation. The modeler designs and implements the ABM, and chooses its parameters and initial conditions (i.e., inputs). Then the model is run to generate an emergent output – if this output can, in some sense, reproduce the phenomenon then it becomes a candidate explanatory model; otherwise it is rejected. Whilst the ABM community is now focusing heavily on methodology for the consideration of model inputs (e.g, calibration techniques), surprisingly little attention is given to the consideration of model structure – i.e., the nature of the entities and equations in the ABM. Whilst initiatives such as the Overview, Design concepts and Details (ODD) Protocol encourage modelers to articulate ABM structure in a thorough manner, these initiatives do not stimulate scientific consideration of the plurality at the heart of model structure selection decisions. Where does a particular structure come from? Does it arise from the art of the modeler, or does it arise from a scientific process? How does ABM speak to theory, and vice versa? How do we choose between alternative mathematical and computational realizations of a specified mechanism, and when do we know that a mechanism can be rejected? What is special about the structure that has been identified, compared to the universe of other structures that could have been chosen? Do multiple, meaningful candidate structures exist and, if so, do these share any similarities? To embark on the journey to answering these questions, the ABM community now needs to bring the issue of model plurality, and methods for model discovery, to the forefront. This workshop will introduce the philosophy and ideas that underpin the concept of model discovery, most recently articulated by Epstein as ‘inverse generative social science’, but which have also arisen under the designations of ‘model crunching’ and ‘structural calibration’; related ideas have also arisen in the areas of pattern oriented modeling, model alignment, and model replication and breaking. The workshop will then set out recent developments in computational intelligence and machine learning methods that have been harnessed for computer-aided model discovery, including how to conduct an efficient search over the wide range of ABM entities and equations that might explain a target phenomenon, and how to use the results to perform abductive inference of key causal mechanisms. The workshop will then introduce recently developed tools that enable these methods to be used by the ABM community – including object-oriented software architectures and synthetic agent populations, NetLogo and Repast implementations, and integrated software platforms for model discovery and inference. Next, the workshop will present learnings from a diverse set of recent applications of model discovery, including civilization growth and decline, alcohol use in US society, and message cascading on social media platforms. Finally, the workshop will set the stage for a discussion on the potential of model discovery as a new grand challenge for the ABM community.

Programme

5-5:20PM
Introduction to inverse generative social science (iGSS) – Joshua Epstein (New York University and Santa Fe Institute, USA)
5:20-5:55PM
New concepts in agent-based model discovery – Ivan Garibay (University of Central Florida, USA) and Robin Purshouse (University of Sheffield, UK) + Panel discussion with Blake LeBaron (Brandeis International Business School, USA) and Doyne Farmer (University of Oxford, UK)
5:55-6:45PM
Applications of iGSS and model discovery: drinking behaviors, population dynamics, segregation, and social media – Talks from Bill Rand (North Carolina State University, USA), Chathika Gunarantne (Massachusetts Institute of Technology and University of Central Florida, USA) and Tuong Manh Vu ((University of Sheffield, UK) + Panel discussion with Blake LeBaron (Brandeis International Business School, USA) and Doyne Farmer (University of Oxford, UK)
6:45-7:00PM
The future of inverse generative social science – Joshua Epstein (New York University and Santa Fe Institute, USA)

<href=”https://www.sheffield.ac.uk/cascade/events/modeldiscovery”>The detailed programme is here.

For info, contact the chair here.