One of the most widely used illustrations regarding the use of SDS modeling concerns management of the deer population on the Kaibab plateau in Arizona. In building a policy that focuses on the predatory population (mountain lions), deer population and forest on the plateau, it is all too easy to end up with a plateau that has no deer, predators or threes. A SDS model shows how this can occur: https://www.bing.com/videos/riverview/relatedvideo?&q=System+Dynamic+Model+Examples&&mid=894CEC12390F6B839F00894CEC12390F6B839F00&&FORM=VRDGAR
In extending the Kaibab analysis to our entire world, it is shocking to witness the death of an entire global system through use of SDS modeling (as engaged by the Club of Rome).
Unfortunately, like DES, the agents in the SDS form of modeling are passive—meaning they operate more like wigits or products than fully functioning decision-makers which we find in the real-world experiences. SDS also lacks the random interjection of unanticipated changes or impacts from unexpected decisions being made as also happens in the real world of healthcare. Patients are not passive products passing through workstations as imagined by early Scientific Method (machine-focused) assembly-lines of the auto-industry and many other manufacturing industries that laid the groundwork for both DES and SDS modeling. Both of these forecasting simulation models are highly sophisticated and can play an important role in moving the efficiency needle of healthcare, yet neither can robustly challenge our broader need of rethinking and re-imagining healthcare that can more broadly invest in the most impactful and cost-effective ways of improving the health of our nation.
Both DES and SDS are sometimes termed “compartmentalized simulations” in that they allow for simulations of contained systems in equilibrium, which can limit the generalizability of their results outside of the studied system. They also do not allow for disequilibrium or allostatic states often found in natural environments where lower-level dynamics can suddenly shift into new unpredictable states due to external forces or accumulation of influences that lead to a tipping point into a higher-level state, such as water going from ice to fluid to steam, when shifted from 32 à 33 degrees and 211 à 212 degrees in very non-linear fashion.
Although both DES and SDS modeling are increasing rapidly in healthcare as we develop more robust data-bases to allow for more sophisticated simulations—we are also seeing a rise in Hybrid Modeling (combining 2 or more simulation models such as DES and SDS) to better simulate real-world healthcare events as well as the emergence of a newer, more effective way to simulate complex systems with active agents who make decisions and interact with other active agents in more complex and random ways, called Agent-Based Simulation (ABS).
Agent-Based Modeling
ABS also emerged as a theory in the 1950’s. Yet, it did not develop into functional simulations until the 1990’s when computer capabilities allowed for it’s more complex and dynamic simulation processes. Its original broad use emerged via the gaming industry in the Game of Life—a complex game that allows for autonomous (active) agents to make decisions that can lead to the expansion of life or death in themselves and their near neighbors. As Nigel Gilbert (2008, p. 1) has noted: “Agent-based simulation has become increasingly popular as a modeling approach . . . because it enables one to build models where individual entities and their interactions are directly represented.”
