Systems Models
Systems models help identify high-leverage points, characterize system modes, communicate insights, and support hypotheses through simulation, enabling better understanding and decision-making in complex systems.
THE IDEA
Identify leverage points
Just as the small movement of the captain's wheel can change the course of a half-million-ton oil tanker, complex systems may contain points of high leverage. System models and simulation tools allow the exploration of a system's sensitivity to changes in various parameters. A large sensitivity to parameter value indicates potential for high leverage. Often, in complex systems, the location of high leverage points may not be intuitive. A model can help reveal interventions that intuition alone might miss. Also, prior system behavior, identified in the course of gathering model data, may give subtle clues to points of high leverage.
Recognize modes
Systems models afford the user the chance to characterize the basic modes of the system. For example, are certain variables exhibiting exponential growth, approach to an asymptote, or perhaps oscillation? Comparing model behaviors to the observed data from the real system can help in understanding the nature of the system relationships. Further, the model may expose potential behavior of the system that had not been anticipated. Thus, the model may help avoid unexpected results.
Communicate
Models are very valuable to communicate and teach. Simulations can be a very impactful tool to illustrate an effect on managers and colleagues. If the simulation can be interactive and provide instant feedback to questions, all the better. Simulation models can also be excellent learning tools when used as management flight simulators. Some of you may have participated in strategic role-play using one of these simulation tools.
The process
Building a systems model can be useful before the simulation is run. Interviews with managers and historical performance data may reveal feedback loops, delays and differences that an aggregate view conceals. Deciding how to represent these relationships also makes assumptions available for discussion.
Generative
System models based on deductive reasoning may be able to provide support for your ideas about why a system is behaving as it is. Scientifically, we would say one is trying to build support for a hypothesis. This has been described by Epstein and Axtell (1996) as generative science. Their argument is that if you can build a simulation using the characteristics of the real world and your simulation produces behavior that is consistent with what you see in the real world, then you have supported, but not proven, your argument about the system mechanisms.
When to use it
A project manager uses a system dynamics model to understand the impact of hiring on project delivery timelines, adjusting parameters to find optimal staffing levels.
A policy analyst simulates the effects of a new regulation on the economy, using the model to communicate potential outcomes to legislators.
An engineer creates a model of a manufacturing process to identify the most sensitive parameters that affect product quality, leading to targeted improvements.
An educator uses an interactive flight simulator to teach students about the feedback loops in ecosystem management, allowing them to explore consequences of different decisions.
FROM THEORY TO EVERYDAY LIFE
In practice
Supply Chain Optimization
Suppose a company’s supply-chain simulation shows that reducing a warehouse inventory buffer improves throughput and lowers costs. This identifies a possible intervention to investigate. The team can examine why it occurs in the model and whether the assumptions fit the warehouse before changing operations.
Policy Impact Simulation
A government agency models a proposed carbon tax. Some simulated scenarios show rapid growth in green technology adoption and emissions approaching a target. Comparing scenarios helps explain the assumptions behind these outcomes and where the policy may fall short.
A MOMENT TO REFLECT
Put it to work.
What potential high-leverage points exist in a complex system you interact with daily, and how could a system model help identify them?
How might building a simulation model of a current project or process reveal unexpected emergent behaviors or modes?
In what ways can a systems model support or challenge your existing hypotheses about how a system works?
References
- Growing Artificial Societies — Epstein and Axtell (1996)
CONTINUE THE CONVERSATION
Source texts
- Thinking in Systems and Mental Models Think Like a Super Thinker (Dawson, Marcus P)