Lecture Z1 (2021-04-20): Final Exam Review

In this lecture, we review topics from the semester in preparation for the upcoming final exam.

Lecture G1 (2021-04-08): Randomness and Chaos

In this lecture, we introduce the concepts of stochastic modeling (the use of randomness to simplify the modeling process) and chaos (the extreme sensitivity to initial conditions that makes deterministic systems appear to be random. To demonstrate stochastic modeling, we develop a discrete event system simulation model of a bacterial population. This gives an opportunity to discuss random number generation/random number streams and how they allow non-random computers to approximate randomness and realistic variation (allowing for realistic experimentation). We pivot to discussing chaos -- using the Mackey-Glass and Lorentz systems as key examples (one that has a single stock with delay in the flow, the other that has three stocks with simple flows). We finish by connecting these results to the so-called "butterfly effect."

Lecture F3 (2021-04-06): Chapter 10, Model Validity, Mental Models, and Learning (Morecroft, 2015)

In this lecture, we review topics brought up in Chapter 10 of Morecroft (2015), which focusses on building confidence in models. In particular, we review the definition of a model and the different types of models -- from analog to metaphorical -- and how the lack of realism in illustrative models can have a strength in being more generalizable. We discuss formal models as learning tools -- transitional objects that help someone move from one mental model to another. We also discuss how formal models help to build communities of knowledge among model builders. This discussion borrows both from Morecroft (2015) as well as work by cognitive psychologist Frank Keil and a few others. We close with a few comments about testing, verification, and validation (TV&V) and a more concrete example.

Lecture F2 (2021-04-01): Chapter 9, Public Sector Applications of Strategic Modelling (Morecroft, 2015)

In this lecture, we highlight and expand on topics covered in Chapter 8 of Morecroft (2015) related to public sector applications of strategic modeling. We first focus on Forrester's Urban Dynamics model, which provides a dynamic hypothesis explaining systemic (endogenous) reasons for growth and eventual stagnation (and even decline) of cities. We then transition to a a more complex fisheries model that includes endogenous investment (i.e., a model where investment decisions are generated by the model as opposed to specified by the person operating the model). This allows us to discuss a more specific definition of "tipping points" phenomena. Ultimately, many related discussions (bifurcation analysis and endogenizing variables) are left for auxiliary videos (for the sake of time).

Lecture F1 (2021-03-25): Chapter 8, Industry Dynamics - Oil Price and Global Oil Producers (Morecroft, 2015)

In this lecture, we review topics covered in Chapter 8 of Morecroft (2015), which is a case study of a system dynamics model of oil price trajectories meant to mimic a variety of periods of history while also helping in strategic scenario planning for the future. This particular model focuses on the effect of the OPEC oil cartel on price dynamics, including both the role of the swing producer in both setting prices and ensuring enforcement of quotas by other OPEC members.

Lecture E5 (2021-03-18): Assignment E5 – Creating Limited, Coupled Population Models

This lecture introduces a more significant homework assignment in SOS 212 that provides practice in building complex, multi-sector models in Vensim and Insight Maker. To aid in completing that assignment, the lecture also provides reminders of how to build equations for rates, how to implement lookup tables, how to use shadow variables (ghost primitives), and how to use the modulo (mod) operator to implement repeated, seasonal patterns.

Lecture E4 (2021-03-16): Chapter 6, The Dynamics of Growth and Diffusion (Morecroft, 2015)

This lecture covers content from Chapter 6 of Morecroft (2015), which focuses on the Bass model of product adoption (innovation diffusion) and related models that assist in decision making about marketing of new products within already crowded markets. Various connections are made between the Bass model (which is effectively a contagion model) and the SIR model covered in Lecture E3.

Lecture E3 (2021-03-11): Epidemic Dynamics

In this lecture, we apply system dynamics modeling to generate basic models of disease spread and scenario planning for intervention strategies in epidemiology and public health. In particular, we introduce the SIR model (and a few variations on it) and discuss the details of how it works and what the behavior over time curves mean. Topics related to current events, such as Ebola, COVID-19, and "flattening the curve", are discussed in relation to these models.

Lecture E2 (2021-03-09): Making Simulations More Realistic, Part 2 - Delays, Fixed and Smoothing

In this lecture, we describe how to implement delays in stock-and-flow diagrams in Vensim and Insight Maker. We focus on two specific types of delays -- fixed delays (as in transport delays) and smoothing/averaging delays (as in exponential decay). The lecture is motivated by a hypothetical exercise of modeling the ingestion process -- where food enters the mouth, travels down the esophagus (fixed delay), and then gets processed in the stomach (smoothing delay) as nutrients move into the bloodstream. Much of this lecture is devoted to building up intuition about what a "delay" parameter is in a smoothing delay -- this parameter captures the inertia (or sluggishness/resistance to change) of systems. That allows us to re-introduce the "time constant" again.

Lecture E1 (2021-03-04): Making Simulations More Realistic, Part 1 - Units, Sliders, and Lookup-Table Converters

In this lecture, we introduce several more advanced techniques in Vensim and Insight Maker that help to make simulations more realistic. We start with the formal use of units and unit conversions in models. We then introduce sliders, which allow for expedited exploration over parameter spaces in simulation models. We close with a demonstration of lookup tables, both for creating converters (specified with graphs or tables) from one variable to another as well as tools for introducing changes over time. Making use of Time in Vensim also requires introducing Shadow Variables.

Lecture D-E (2021-02-23): Midterm Review

This lecture serves as a review of the pre-midterm material in SOS 212 (Systems, Dynamics, and Sustainability) in the Spring 2021 semester at Arizona State University. It reviews basic topics on modeling philosophy, causal loop diagrams, systems archetypes, and basic stock-and-flow diagram/simulation concepts.

Lecture D4 (2021-02-18): Chapter 3, Modelling Dynamic Systems (Morecroft, 2015)

In this lecture, we start with a brief tutorial of simulating simple stock-and-flow diagrams (Systems Dynamics Models) in Vensim and Insight Maker. We then discuss Chapter 3 (Modelling Dynamic Systems) from Morecroft (2015), which introduces building and simulating dynamical systems models.

Lecture D3 (2021-02-16): Stock and Flow Diagrams in Vensim and Insight Maker

This lecture continues our introduction to stock-and-flow diagrams, moving from implementation in spreadsheets to implementation in special-purpose simulation software such as Vensim and Insight Maker. We also cover the motivation behind common flow expressions for population-growth systems and systems that fill or empty over time (which allows us to define "time constant").

Lecture D2 (2021-02-11): Introduction to Numerical Simulation of Dynamical Systems, Part 2

This lecture continues where the previous one left off in the introduction of numerical simulation of dynamical systems. We spend most of the lecture working through the bacterial growth example and adding a death process. This allows us to better highlight the ways that stocks and different flows can interact. We conclude with a challenge to simulate a toilet-tank example in a spreadsheet, which is a negative feedback loop that very different "flow" formulas but otherwise is simulated in a very similar way to the bacterial growth case.

Lecture D1 (2021-02-09) - Introduction to Numerical Simulation of Dynamical Systems, Part 1

In this lecture, we more formally introduce the dynamical processes that drive system behaviors over time. To motivate how these dynamical processes can be simulated, we consider two different ways to calculate compound interest in a bank. The second way, which involves calculating how much interest is generated each year, is identical to how we numerically integrate dynamical processes captured as stock and flow diagrams. We discuss a simple bacterial growth modeling problem as an example.

Lecture C2 (2021-02-04): "Applying Systems Archetypes" (Kim and Lannon, 1997)

In this lecture, we review the article by Kim and Lannon (1997) on "Applying Systems Archetypes." This article introduces System Archetypes, a set of combinations of feedback loops that common occur in systems and are associated with a set of problems, possible solutions, and likely behaviors over time (BOT). This article portrays the System Archetypes as tools that can be used in one of four different ways -- as structural pattern templates, as lenses for highlighting subtle but important potential aspects of problems, as dynamic theories for understanding and invention, and as tools for predicting future behavior based only on the internal components of a system. We explore these four different ways of applying the archetypes and tie them in to the scientific method we practice within Sustainability Science which is often assisted through the use of computer simulation modeling.

Lecture C1 (2021-02-02): Feedback Systems Thinking with CLDs

In this lecture, we introduce ways to build up more complex thinking using Causal Loop Diagrams by introducing fundamental modes of dynamic behavior. These fundamental modes can generate aspects of many common behaviors over time just through the combination of a few patterns of feedback loops and delays. This lecture helps motivate the more complex System Archetypes that will be introduced in the next lecture.

Lecture B3 (2021-01-28): Chapter 2, Introduction to Feedback Systems Thinking (Morecroft, 2015)

In this lecture, we discuss Chapter 2 from Morecroft (2015), which introduces feedback systems thinking. Morecroft contrasts feedback systems thinking with an event-oriented world view. To aid in taking the systems perspective, causal loop diagrams (CLDs) are introduced with connections to the fundamental modes of behavior through time that are associated with common causal structures. Examples are given in the chapter from a number of different application spaces, but the lecture focuses on examples from road construction and manufacturing.

Lecture B2 (2021-01-26): Causal Loop Diagrams in Vensim

In this lecture, we start drawing Causal Loop Diagrams (introduced in the previous lecture) in the System Dynamics Modeling tool Vensim (from Ventana Systems). This serves as a brief introduction to Vensim without getting into the stock-and-flow diagrams that we will learn to draw, simulate, and analyze later in the semester. This lecture also provides more practice in drawing CLD's, which help to diagram the causal structure in complex systems and identify the possibly many feedback loops that enrich the dynamics of such systems.

Lecture B1 (2021-01-21): Introduction to Causal Loop Diagrams

In this lecture, we move from mental/qualitative models toward building quantitative modeling tools to do strategic modeling and scenario planning, as described in Chapter 1 of Morecroft (2015). This involves giving a preview of stock-and-flow diagrams (introducing stocks and flows) and then moving into introducing causal loop diagrams (CLD's). Causal loop diagrams provide a graphical accounting of all causal connections in a model of a system and help to identify both negative and positive feedback loops that are informative about likely dynamical modes of behavior.

Lecture A3 (2021-01-19): Chapter 1, The Appeal and Power of Strategic Modeling (Morecroft, 2015)

In this lecture, we cover topics from Chapter 1 (The Appeal and Power of Strategic Modeling) by Morecroft (2015). This chapter helps to motivate the use of quantitative, simulation models for learning, as opposed to just prediction. It uses a board game (Monopoly) to motivate how overly simplistic models can nevertheless provide more generalizable insights, which is one of the reasons why computer simulation models can still be very useful even if they do not include every detail. We close with a simple fisheries example that helps to underscore the importance of choosing the correct functional response (type-I or type-II, for example) when modeling predator--prey dynamics.

Lecture A2 (2021-01-14): Introduction to Modeling

In this lecture, we build upon the basic definition of a model as a tool for answering a "What If" question. We point out that there are mental (qualitative) models as well as quantitative models, and we need to learn how to recognize our internal mental models. We then discuss the benefits of the different modeling types and setup the foundation for entering into more quantitative modeling approaches that are focused on in the rest of the class.

Lecture A1 (2021-01-12): Course Introduction

In this lecture, SOS 212 (Systems, Dynamics, and Sustainability) is introduced to Spring 2021 semester students at Arizona State University. The basic learning outcomes for the course are presented as well as the course structure and policies.