Learning from DMDU and Complexity

By Robert Lempert

Those seeking to make decisions or to exercise leadership in the world confront uncertainty because the future is not simply an extension of the past. The concept of deep uncertainty refers to situations “when parties to a decision do not know, or cannot agree on, the system model that relates action to consequences, the probability distributions to place over the inputs to these models, which consequences to consider and their relative importance. Deep uncertainty often involves decisions that are made over time in dynamic interaction with the system.”[1]

Principles of Decision Making under Deep Uncertainty (DMDU) include considering multiple futures, not just one, in planning; seeking “robust” plans that perform well over many futures, instead of optimal plans designed for a single, best-estimate future; and making plans flexible and adaptive. Common tools include expert convening, decision-structuring, exploratory modeling, designing policy alternatives, and developing and stress testing adaptive strategies for implementing policy.

Deep uncertainty is also an attribute of complex adaptive systems (CAS), which are characterized by having many interconnected, interdependent parts that change behavior as a system and surrounding context change over time. For policy makers and leaders managing a complex system is fundamentally different than managing a complicated one. With a complicated system, decision-makers seek to predict and then control its behavior. With a complex system there is often a much wider range of actors, and those making decisions seek to understand the internal logic of a system and its different contingent pathways, probe to understand a system’s current state, and respond.

This panel will explore how DMDU methods and tools can address complex policy challenges from climate change to healthcare. The panelists explore how tools and frameworks from modeling, decision-making, and leadership can relate and enrich our ability to engage complex policy challenges. The speakers discuss the tools through a range of applications, and cover topics from analyzing behaviors and adaptations of systems, to methods for reshaping system structures.

The session, chaired by Robert Lempert from the RAND Corporation, will have four talks:

  • Computational Intelligence for Studying Sustainability Challenges: Tools and Methods for Dealing with Deep Uncertainty and Complexity.Edmundo Molina-Perez, Tech de Monterrey
  • Classification in the Wild: The Science and Art of Transparent Decision MakingKonstantinos Katsikopoulos, Southampton Business School
  • Using ABMs to Understand the Responses of Ocean Fisheries and Fishers to Climate ChangeRobert Axtell, George Mason University and Santa Fe Institute
  • What Can We Learn from Mexican Healthcare Reform for Transformation of Large, Complex Systems?Tim McDonald, Pardee RAND Graduate School

Examining sustainability challenges, Molina-Perez shows that the combination of high-performance computing, mathematical modeling, and computational intelligence tools such as optimization and clustering algorithms can lead to richer analytical insights. He proposes an analytical hierarchy of computational tools that can be applied to other sustainability challenges.

Katsikopoulos describes an approach to improving decision-making theory to reflect conditions under realistic conditions of uncertainty – or, “in the wild” – that are not captureed in typical psychological experiments or decision-theoretic models. He introduces heuristics to help practitioners make decisions while also testing how to improve decisions in an adaptive way.

Axtell will show how simple models can be policy relevant for complex systems, highlighting the use of agent-based models (ABM) for the experience of planning responses of ocean fisheries to climate change.

McDonald describes principles of leadership for changing large, complex systems under conditions of deep uncertainty, using case study analysis of the transformation of a national health system that substantially expanded health coverage for its previously uninsured population.

Decisionmakers need new ways to grapple with deep uncertainty and complexity. Such decision-making requires being explicit about goals, considering a range of alternative options and tradeoffs, listening to the input of others, using the best available science and evidence to understand the potential consequences, and following norms that support the legitimacy of the decision-making process. Together these presentations explore issues of analysis, planning, and strategic decision making in complex environments. While managing a complex system in many ways proves more challenging than managing a complicated one, it also offers more opportunities because there are more potential points of engagement, and small interventions can sometimes make large differences in the state that emerges. The key question, taken up by this panel, is how this can be done and what can be learned from DMDU and complexity.

[1] http://www.deepuncertainty.org/


About this Blog Post:

This blog post is part of a series of posts contributed by the chairs of the 2020 DMDU Annual Meeting. For more information about the Annual Meeting, including registration, visit our website at 2020.deepuncertainty.org. We hope to (virtually) see you soon!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.