# 2.3.1: Strengths of ABM Having seen what agent-based models are made of, it is worth pausing to ask what they are good for, and why anyone would choose this approach over a simpler one. The strengths of ABM follow directly from the way it is built: because it represents individuals, it can hold their differences, place them in space or a network, let them adapt, and reveal how patterns at the level of the whole grow out of local behaviour. Knowing these strengths matters because they are the reasons to reach for ABM in the first place, and because each one hints at the kind of question the method answers best. ## 2.3.1.1: The eight strengths at a glance Eight recurring strengths make ABM worth the extra effort of modelling individuals rather than averages. Each card below names one and gives a one-line summary; click a card to read what it means and why it is useful, and notice how several come back to the same core idea, that modelling individuals lets you see things an aggregate model cannot. ```{raw} html :file: ../learning_activities_and_widgets/widgets/widget_2_3_1a_strengths_of_abm.html ``` ## 2.3.1.2: Why the strengths cohere The strengths of ABM nearly all flow from one choice, modelling individuals rather than averages, which is what lets it capture heterogeneity, space, adaptation, and the link from local behaviour to system-level patterns. The next two examples make this concrete: the first drills into a single strength, emergence, and the second shows several strengths working together in one model. Before those examples, confirm the through-line for yourself: every strength traces back to the one choice of modelling individuals. ```{raw} html :file: ../learning_activities_and_widgets/learning_activities/activity_2_3_1b_one_choice_many_strengths.html ``` ## 2.3.1.3: Strengths in action: emergence Emergence is the strength that most rewards a worked example, because its whole point is that the system-level pattern is nowhere written in the individual rules. The following case, video and hands-on activity all illustrate it through crowd evacuation. ```{prf:example} Emergent behaviour: evacuation dynamics :label: ex-evacuation-dynamics Evacuation simulations provide a clear example of emergent behaviour. When individual agents (people) each follow simple, rational rules, such as moving towards the exit, complex crowd behaviours nonetheless emerge naturally. Bottlenecks and congestion form at exits even when everyone tries to move efficiently, and counter-intuitive patterns arise, such as the faster-is-slower effect, in which moving too quickly can actually slow overall evacuation. As shown in the experiment video, when all agents try to pass through a door at once, individually rational behaviour (rushing towards the exit) produces collectively irrational outcomes (jamming and slower evacuation). This demonstrates how ABM reveals the gap between individual optimality and system optimality, an insight that would be impossible to obtain through equation-based models or intuition alone. ```
Experiment showing emergent behaviour in crowd dynamics and evacuation scenarios.
The video shows the effect once; the activity below lets you produce it yourself. Here a crowd leaves a room through one exit, and you control how much urgency drives them to rush and push. Before you run it, predict: does rushing empty the room faster? ```{raw} html :file: ../learning_activities_and_widgets/learning_activities/activity_2_3_1c_faster_is_slower.html ``` Intuition can get a system badly wrong, as with the faster-is-slower effect, and one of the real strengths of ABM is that by simulating the individuals it can surface such counter-intuitive results before they surprise us in reality. ## 2.3.1.4: Strengths in action: a whole model A single strength is easiest to see in isolation, but a real model usually draws on several at once. An urban planning ABM demonstrates how the strengths combine in one place: ```{admonition} Example: urban planning ABM benefits An urban planning ABM demonstrates these strengths: - **Natural representation**: Individual households, businesses, and developers - **Emergent behaviour**: Urban sprawl patterns emerge from individual location choices - **Policy testing**: Test zoning changes, tax policies, transport investments - **Heterogeneity**: Different types of residents with varying preferences - **Spatial effects**: Location, accessibility, and neighbourhood effects matter - **Stakeholder engagement**: City planners can visualise and discuss model scenarios ``` The model above lists its strengths together; the activity below asks you to place each one where it appears. ```{raw} html :file: ../learning_activities_and_widgets/learning_activities/activity_2_3_1d_spot_the_strength_at_work.html ``` Taken together, these examples answer the question the page opened with: ABM earns its place whenever the individuals, their differences, their arrangement in space or a network, and the patterns those interactions produce are the very things that must be understood. Every strength on this page is a facet of that one advantage. The same commitment to modelling individuals also carries costs, and the next section turns to the limitations that come with it. ## Key takeaways ```{admonition} Key takeaways :class: important - The strengths of ABM follow from a single design choice: representing individuals rather than population averages. - That choice allows a model to capture heterogeneity, spatial and network structure, and adaptation, to link individual behaviour to system-level patterns, and to serve as a virtual laboratory for testing policies. - Because it simulates individuals, ABM can reveal counter-intuitive outcomes, such as the "faster is slower" evacuation effect, that aggregate models cannot. ```