Some examples of ABS analyses resemble those used by SDS in its modeling of the Kabab plateau eco-system. Specifically, ABS has been used to model Wolves, Sheep and Grass. This model illustrates a dynamic interplay of populations based on initial population size and reproduction rates. It allows for very realistic predictions of population dynamics in US parks: https://ccl.northwestern.edu/netlogo/models/WolfSheepPredation

Compared to our prior two models, ABS has an attractive array of factors—beyond just the representation of individual actors–that increases its capacity to simulate reality in healthcare:

1. Stochastic—allows for random events and decisions to impact outcomes
2. Active Agents—agents are active decision-makers, not simple products or passive people flowing through a system.
3. Feedback loops—autonomous, active agents interact with each other within the context of the simulation, producing both immediate and delayed impacts on outcomes of the system.
4. Emergence—allows for emergent behaviors to impact broader environment of the simulation.

Most agent-based models are composed of numerous agents that exist at various scales. The agents might be individuals, groups, organizations or entire communities. These models also require some mode of decision-making—usually a mode that is rather simple and straight-forward. A third requirement of an ABM is that it include some process for bringing about change in the behavior of participating agents. Learning and adaptation occur that lead to the modeling of diffusion of ideas and to the tipping points made “famous” by Malcolm Gladwell (2000).  Fourth, an agent-based model requires a “field” (topology) on which the agents interact with one another. As in the case of SDS, boundaries are established so that an analysis can be conducted that focuses on the behavior of agents operating in this specific field.

All of this is done so that a computer-based simulation can be engaged to determine ways in which the behavior of a single individual can impact on the overall behavior of agents operating in a specific system. Unlike DES and SDS, agent-based modeling opens the door for exploration of the critical, but often elusive, dynamic of Emergence. It is in the capacity of a complex, adaptive system to produce something that is emergent (absolutely new and surprising) that this system is ultimately most adaptive and capable of higher order learning and change.

Confronting the Virus on a Warped Plane

System dynamics modeling moves us past the static and mechanistic world of DES. It provides us with a portrait – a “snapshot”—of a system that is complex, with many interrelated parts. Agent-based modeling, in turn, moves us to a computer-based narrative—a “move”–of a dynamic, interacting system. However, neither of these modeling tools conveys something of what it is like to actually navigate through a complex, adaptive system. What is it like from moment to moment when moving through a field (typology) that is dancing—such as in our contemporary world of health care. How might the life of a leader in health care be portrayed on a daily basis—especially when this leader is confronting a powerful, complex challenge such as COVID-19.