layout: true <div class="my-footer"><img src="../img/presentation/segue23/bayili/TUBK_Logo_URBA.svg" alt></div> --- <!-- About Agent-Based Modelling: Layers --> .right[.image-90[]] ## Agent-Based Modelling ABM is a bottom-up approach to modelling, simulating collective social phenomena from micro behaviours. ### Entities In ABM, there are typically two layers of entities: the actor layer and the environment layer: - Actors The human actors, <u>active agents</u>, can move around and interact with other entities. - Environment The environment, <u>passive agents</u>, is where the active agents reside. --- <!-- About Agent-Based Modelling: Rules and interactions --> .right[.image-90[]] ## Agent-Based Modelling ### Rules - The rules guides how the agents behave and interact with each other. - These rules are based on both dynamic and static attributes. ### Interactions - Actors interact with each other and the environment, altering their states. - The interactions can vary over time and space. --- <!-- About Agent-Based Modelling --> .right[.image-90[]] ## Agent-Based Modelling ### Emergent Phenomena - The actors' and environment's changing states indicate <u>social and environmental impacts</u> on the system. - The social and environmental impacts contribute to <u>the phenomena of the system</u> collectively. In form of **urban economic segregation**: - What are the key entities? - What are the key behaviour rules? --- <!-- The overview of the ABM's theoretical framework --> # Overview ## The target social phenomena **Urban economic segregation** > The uneven distribution of groups with **different economic status** across **residential neighbourhoods** of a city or an urban region (adapted from Van Ham et al. (2021)). Two fundamental questions: - Why do people choose to live where they live? - How the their economic status change over time? ### Research question How can agent-based modelling enhance our understanding of the <u>dynamics driving urban economic segregation patterns</u> through the study of **individual residential and economic behaviours and interactions** in cities? --- .right[ .caption[.center[Theoretical framework from micro behaviours <br>to social phenomena through ABM]]] # Overview ## Purpose - To create a versatile and adaptable ABM framework - To model the intricate interplay of residential and economic behaviours - To provide simulated evidence for policymakers. --- # Review .right[ .caption[.center[Collection of the related ABMs from the literature]]] Residential mobility is well-modelled due to its relevance in wide range of social phenomena. --- # Review .right[ .caption[.center[Collection of the related ABMs from the literature]]] Residential mobility is well-modelled due to its relevance in wide range of social phenomena. Economic mobility is less modelled, especially in spatial contexts. --- # Review .right[ .caption[.center[Collection of the related ABMs from the literature]]] Residential mobility is well-modelled due to its relevance in wide range of social phenomena. Economic mobility is less modelled, especially in spatial contexts. ABMs of segregation primarily discussed location in segregation but not the complexity of residential choice or economic behaviour. --- # Review .right[ .caption[.center[Collection of the related ABMs from the literature]]] Rather small number of ABMs captured the residential mobility and economic mobility simultaneously in the same segregation model. --- # Review ## The key actors - **Household** Household is the most common actor modelled in related ABM. - **Individual** Individual, may called agent, is commonly modelled as the member of household. Depending on the social phenomena and mechanism selected, the individual agent can be substituted by other actors. - **Other actors** The other actors, developer, urban sector or government, etc, are modelled as actor of interests in some ABMs which involves dynamic planning process. --- .right[] ## The key components **Life Course Component** The life course regulates the internal attributes of the agent, posing event and transition as a trigger of further responses on residential and economic mobility. --- .right[] ## The key components **Residential Mobility Component** The household make their residential decision, the residence location change can impact the social network change and the economic states. --- .right[] ## The key components **Economic Mobility Component** The economic states change can impact the residential decision. --- .right[] ## The key components **Social Network Component** Social networks serve as a foundational mechanism for capturing the complexities of social influence on behaviour. --- .right[] ## The key components **Macro Level Regulation** The macro level regulation was conceptualised as the exogenous dynamics. It can impact the economic states and the residential decision. It imitates the role of policy interventions, and will be used to create the policy scenarios. --- .right[] # Review ## Mechanisms and processes Green box: widely included Red box: not well modelled --- .right[] # Review ## The knowledge base The location-based opportunity is well captured in the ABMs of urban migration by uneven development or specific function-oriented location choice such as job-oriented or education-oriented. --- .right[] # Review ## Gap The tenure choice in residential mobility was widely researched theoretically, but rarely modelled together in ABM. --- <!-- Prototype Model design --> ## Prototype Model design ### Case study .image-40[] .caption[.center[Figure. The case of Groningen and its population density.]] The [Functional Urban Areas (FUA)^a](https://www.oecd.org/cfe/regionaldevelopment/functional-urban-areas.htm) of Groningen. The population of FUA of Groningen is <u>194, 340</u> in 2011 census, aggregated to around <u>80, 000</u> households. .note[*Notes: ^a: each of FUA consists of a densely inhabited city and of a surrounding area (commuting zone)*] --- <!-- The life course component --> ### The life course component .right[ .caption[.center[Figure. An example of life trajectory, events and impact on household characteristics.]]] **Fundamental stages and essential micro life events** The trajectories of individuals and households are intertwined with each other through the life events. --- <!-- The life course component --> ### The life course component .right[ .caption[.center[Figure. An example of life trajectory, events and impact on household characteristics.]]] **Fundamental stages and essential micro life events** Life events have the potential to trigger relocation desires, influence preferences for residence and economic states. --- <!-- Residential mobility component --> .right[] ### Residential mobility component #### The residential utility Residential Utility is defined by static attributes and dynamic preference varying over households. The factors of perceived utility of housing include the accessibility to services, the living unit size fits the household size and the social group by income and demographic background^a. .note[*Notes: ^a: The preference level of residing with similar economic and demographic neighbours.*] --- .right[] #### The housing decision process <!-- preference, expectation, affordability, decision --> Rules: The household will start the housing searching process if the utility of the current housing is lower than the expected utility. --- .right[] #### The housing decision process <!-- preference, expectation, affordability, decision --> Rules: Household with higher disposable income on housing has <u>the priority</u> to select the housing units. --- .right[] #### The housing decision process <!-- preference, expectation, affordability, decision --> Rules: The household will <u>increase the searching radius</u> if there is no qualified housing unit within the current searching radius. --- <!-- Entities introduction starts here --> ### Entities (**change it more design related**) .right[] #### Individuals The individuals are the smallest active agent. -- The dynamic choice attributes are updated based on the former experience. The dynamic state attributes are the properties or characteristics of agents that can change over time as the simulation progresses. --- ### Entities (**change it more design related**) .right[] #### Household The individuals are the members of the household^a, and the household is the minimum relocation decision making unit of the model. .note[*Notes: ^a The household covers the family and non-family household. The family household is the household with at least two members, and the non-family household is the household with one adult member.*] --- ### Entities (**change it more design related**) .right[ ] #### Living Unit The living units are the housing units in the grids of residence. --- ### Entities (**change it more design related**) .right[ ] #### Spatial Unit The spatial units are the environment where the active agents live in. In our case, they are the grids of the city. <!-- Entities introduction ends here --> --- ## Limitations of the prototype model to be developed ### Decision rule Some of rules are largely simplified, for example, the behaviour in housing market. ### The tolerance of the decision At circumstances of certain events, such as job change or divorce, the tolerance of the residential decision will be increased. ### Absence of migration flow in and out of the system The long-distance migration flow in and out of the system is not included in the prototype model. --- ## Overall challenges ### Diversity of individuals Without the diversity of individuals, the model might fail to capture endogenous dynamics of the main actors. ### The balance between the complexity and the simplicity of the model The more elements and the more detailed the behaviour and interaction rules, the larger the model needs to be and the longer the simulation will take. ### The interpretability of the model As the intercorrelation between different mechanisms, the impact of each mechanism is difficult to be isolated. Sensitivity analysis and modelling them in building block will be the key to address this issue. --- # Evaluation **What is the ideal ABM of urban economic segregation we are looking for?** - A general decision making mindset - A framework that simulated reliable micro behaviours - A flexible and extendable framework - A representative model of the target social phenomena with essential actors and mechanisms --- **References** <div class="references"> [1] Ma, Y., Zhenjiang, S.H.E.N. and Kawakami, M., 2013. Agent-based simulation of residential promoting policy effects on downtown revitalization. Journal of Artificial Societies and Social Simulation, 16(2), p.2. <br> [2] Pagani, A., Ballestrazzi, F. and Binder, C.R., 2022, March. Shrinking Housing’s Size: Using Agent-Based Modelling to Explore Measures for a Reduction of Floor Area Per Capita. In Advances in Social Simulation: Proceedings of the 16th Social Simulation Conference, 20–24 September 2021 (pp. 275-287). Cham: Springer International Publishing. <br> [3] Clark, W.A. and Onaka, J.L., 1983. Life cycle and housing adjustment as explanations of residential mobility. Urban studies, 20(1), pp.47-57. <br> [4] Chatterjee, K., Sherwin, H. and Jain, J., 2013. Triggers for changes in cycling: the role of life events and modifications to the external environment. Journal of transport geography, 30, pp.183-193. <br> [5] Tabata, M. and Eshima, N., 2003. A self-referential agent-based model that consists of a large number of agents moving stochastically in a discrete bounded domain. Applied mathematics and computation, 143(2-3), pp.443-483. <br> [6] Cui, J., Pan, Q., Qian, Q., He, M. and Sun, Q., 2013. A multi-agent dynamic model based on different kinds of bequests. Physica A: Statistical Mechanics and its Applications, 392(6), pp.1393-1397. <br> [7] Haase, D., Lautenbach, S. and Seppelt, R., 2010. Modelling and simulating residential mobility in a shrinking city using an agent-based approach. Environmental Modelling & Software, 25(10), pp.1225-1240. <br> [8] Acheampong, R.A. and Asabere, S.B., 2021. Simulating the co-emergence of urban spatial structure and commute patterns in an African metropolis: A geospatial agent-based model. Habitat International, 110, p.102343. <br> [9] Jordan, R., Birkin, M. and Evans, A., 2011. Agent-based modelling of residential mobility, housing choice and regeneration. In Agent-based models of geographical systems (pp. 511-524). Dordrecht: Springer Netherlands. <br> [10] Gil-Quijano, J., 2008. Mechanisms of automated formation and evolution of social-groups: A multi-agent system to model the intra-urban mobilities of Bogotá city. In Social Simulation: Technologies, Advances and New Discoveries (pp. 151-168). IGI Global. <br> [11] Gonzalez-Estevez, J., Cosenza, M.G., López-Ruiz, R. and Sanchez, J.R., 2008. Pareto and Boltzmann–Gibbs behaviours in a deterministic multi-agent system. Physica A: Statistical Mechanics and its Applications, 387(18), pp.4637-4642. <br> [12] Pagani, A., Ballestrazzi, F., Massaro, E. and Binder, C.R., 2022. ReMoTe-S. Residential mobility of tenants in Switzerland: an agent-based model. Journal of Artificial Societies and Social Simulation, 25(2). <br> [13] Silveira, J.J., Espíndola, A.L. and Penna, T.J.P., 2006. Agent-based model to rural–urban migration analysis. Physica A: Statistical Mechanics and its Applications, 364, pp.445-456. <br> [14] Sun, S. and Manson, S.M., 2015. Simple agents, complex emergent city: agent-based modelling of intraurban migration. Computational approaches for urban environments, pp.123-147. <br> [15] Boulahbel-Bachari, S. and El Saadi, N., 2022. Understanding the Labor Market from the Bottom Up with an Agent-Based Model. In Proceedings of the Computational Methods in Systems and Software (pp. 754-769). Cham: Springer International Publishing. <br> [16] Taniguchi, T., Takahashi, Y. and Nishikawa, I., 2010, August. Multi-agent simulation on relationship between individuals' travel behaviour and residential choice behaviour. In Proceedings of SICE Annual Conference 2010 (pp. 1728-1733). IEEE. <br> [17] Fatmi, M.R. and Habib, M.A., 2017. Baseline synthesis and microsimulation of life-stage transitions within an agent-based integrated urban model. Procedia Computer Science, 109, pp.608-615. <br> [18] Haase, D., Lautenbach, S. and Seppelt, R., 2010. Modeling and simulating residential mobility in a shrinking city using an agent-based approach. Environmental Modelling & Software, 25(10), pp.1225-1240. <br> [19] Shen, Z., Ma, Y., Kawakami, M. and Nishino, T., 2012. Development of Agent-Based Model for Simulation on Residential Mobility Affected by Downtown Regeneration Policy. In Intelligent Interactive Multimedia: Systems and Services: Proceedings of the 5th International Conference on Intelligent Interactive Multimedia Systems and Services (IIMSS 2012) (pp. 201-211). Springer Berlin Heidelberg. <br> [20] Buchmann, C.M. and Schwarz, N., 2014. Residential mobility and model complexity-an agent-based modelling experiment on a small shrinking town in Eastern Germany. <br> [21] Rosenfield, A., Chingcuanco, F. and Miller, E.J., 2013. Agent-based housing market microsimulation for integrated land use, transportation, environment model system. Procedia Computer Science, 19, pp.841-846. </div>