SDA video modules

This page contains a series of short video modules that started as a three-hour tutorial for the 2026 AIMOR workshop in Banff. Each module is 5 to 25 minutes long, and lists guidelines on technical level and recommended predecessor modules.

The initial modules were posted in August, 2026. I anticipate improving the presentations and adding new modules, so pay attention to the posting and updating dates.

If any faculty would like access to the PowerPoint slides for use in lectures, email me at wbpowell328@gmail.com.

SDA video module thumbnail

My background

Summary: Experience matters, and in this brief presentation I give a peek into my 40+ years of experience thinking about, working on, and implementing systems for solving a wide range of sequential decision problems.

Predecessor modules: None
Technical level: Non-technical
Time: 9:50  ·  Posted: August 22, 2026

SDA video module thumbnail

History of research on sequential decision problems

Summary: This is a brief history starting with "reinforcement learning" circa 1900, up through its re-introduction in the 1950s, optimal control in the 1970s, modern reinforcement learning in the 1980s, and approximate dynamic programming in the 1990s.

Predecessor modules: None
Technical level: None
Time: 21:12  ·  Posted: August 22, 2026

SDA video module thumbnail

Introduction to sequential decision problems

Summary: A presentation of examples of the diversity of sequential decision problems, from the simple inventory problems to complex logistics problems, including an initial introduction to optimal learning problems.

Predecessor modules: None
Technical level: None
Time: 7:19  ·  Posted: August 22, 2026

SDA video module thumbnail

Framing decision problems

Summary: This is the first step to understanding a sequential decision problem, which starts by identifying metrics, decisions and uncertainties.

Predecessor modules: None
Technical level: None
Time: 25:20  ·  Posted: August 22, 2026

SDA video module thumbnail

The universal modeling framework (high level)

Summary: This is how we model any sequential decision problem, consisting of five elements.

Predecessor modules: Framing decision problems
Technical level: Basic modeling notation
Time: 11:18  ·  Posted: August 22, 2026

SDA video module thumbnail

Estimation methods

Summary: Modeling and solving sequential decision problems draws heavily on the use of estimation methods from machine learning. This is a high level overview.

Predecessor modules: Universal modeling framework
Technical level: Basic mathematics of probability and statistics
Time: 20:11  ·  Posted: August 22, 2026

SDA video module thumbnail

Universal modeling framework (detailed)

Summary: I now step through the individual elements of the universal modeling framework.

Predecessor modules: The universal modeling framework (high level)
Technical level: Expectations of simulations of policies
Time: 31:57  ·  Posted: August 22, 2026

SDA video module thumbnail

State variables

Summary: This module addresses the astonishing lack of understanding of state variables in various communities (applied probability, dynamic programming). State variables are fundamental to sequential decision problems.

Predecessor modules: Universal modeling framework (detailed)
Technical level: Some modest mathematics
Time: 28:19  ·  Posted: August 22, 2026

SDA video module thumbnail

Designing policies

Summary: This is the first introduction to the four classes of policies.

Predecessor modules: The universal modeling framework (high level)
Technical level: Low
Time: 15:29  ·  Posted: August 22, 2026

SDA video module thumbnail

Choosing policies

Summary: Having presented all four classes of policies, we now tackle the question of actually choosing one (or a hybrid).

Predecessor modules: Designing policies
Technical level: Moderate
Time: 9:58  ·  Posted: August 22, 2026

SDA video module thumbnail

Policy function approximations

Summary: These are analytical functions that map states to actions, without an imbedded optimization problem.

Predecessor modules: Designing policies, Choosing policies
Technical level: Modest
Time: 9:43  ·  Posted: August 22, 2026

SDA video module thumbnail

Cost function approximations

Summary: We introduce the powerful idea of using parameterized versions of deterministic optimization models that are tuned to work well over time.

Predecessor modules: Policy function approximations; The universal modeling framework
Technical level: Uses deterministic optimization, tuned to work well in a simulated environment
Time: 10:08  ·  Posted: August 22, 2026

SDA video module thumbnail

Value function approximations I — Introduction to dynamic programming

Summary: We start our discussion by introducing Bellman's equation and describing the three curses of dimensionality.

Predecessor modules: The universal modeling framework; State variables
Technical level: Moderate-advanced
Time: 13:19  ·  Posted: August 22, 2026

SDA video module thumbnail

Value function approximations II — VFA-based policies

Summary: We introduce the idea of creating a policy that uses an approximation of the state that a decision takes us to.

Predecessor modules: VFA I; State variables
Technical level: Moderate-advanced — illustrates mathematics with visuals
Time: 20:59  ·  Posted: August 22, 2026

SDA video module thumbnail

Value function approximations III — Reinforcement learning

Summary: We introduce approximate dynamic programming as it is typically presented under the banner of "reinforcement learning," using the language typically found in computer science.

Predecessor modules: VFA II
Technical level: Moderate — math at the level of Sutton and Barto
Time: 17:51  ·  Posted: August 22, 2026

SDA video module thumbnail

Value function approximations IV

Summary: We present the classical techniques of forward approximate dynamic programming, with an emphasis on the use of linear models to approximate the value function.

Predecessor modules: VFA I–III
Technical level: Advanced
Under development

SDA video module thumbnail

Value function approximations V — Applications

Summary: This segment summarizes a series of applications using VFA-based policies, typically in the context of resource allocation problems.

Predecessor modules: TBD
Technical level: TBD
Under development

SDA video module thumbnail

Direct lookahead approximations I — Overview

Summary: We set up the notation for lookahead models, and identify the different ways to approximate lookahead models to make a decision now.

Predecessor modules: The universal modeling framework
Technical level: Advanced
Time: 8:25  ·  Posted: August 22, 2026

SDA video module thumbnail

Direct lookahead approximations II — Deterministic lookaheads

Summary: A generalization of cost function approximations (which do not plan into the future) is to parameterize deterministic models which do plan into the future.

Predecessor modules: DLA I; Cost function approximations
Technical level: Understanding of deterministic optimization
Time: 6:25  ·  Posted: August 22, 2026

SDA video module thumbnail

Direct lookahead approximations III — Stochastic lookaheads

Summary: Focuses on the issue of choosing the "policy-within-a-policy" required by stochastic lookahead models. We illustrate PFAs, CFAs and VFAs.

Predecessor modules: Policy function approximations; Cost function approximations; Value function approximations
Technical level: Advanced
Time: 15:03  ·  Posted: August 22, 2026

SDA video module thumbnail

Direct lookahead approximations IV — Discrete actions

Summary: We introduce the idea of using decision trees for small problems, and Monte Carlo tree search for larger problems.

Predecessor modules: DLA I
Technical level: Moderate
Time: 16:37  ·  Posted: August 22, 2026

SDA video module thumbnail

Direct lookahead approximations V — Vector-valued actions

Summary: We introduce two-stage stochastic programming illustrated by its use in power planning. We then introduce the idea of solving the stochastic lookahead problem using classical stochastic search.

Predecessor modules: DLA I, DLA IV
Technical level: Advanced
Time: 9:46  ·  Posted: August 22, 2026

SDA video module thumbnail

Policy search

Summary: We address the challenge of tuning a parameterized policy using derivative-based and derivative-free stochastic search. Stochastic search is presented as a sequential decision problem, and illustrated using a mutual fund cash balance problem.

Predecessor modules: Universal modeling framework; Policy function approximations; Cost function approximations; DLA II
Technical level: Moderately advanced
Time: 15:10  ·  Posted: August 22, 2026

SDA video module thumbnail

Optimal learning

Summary: One of the most important sequential decision problems involves learning over time, either in an offline setting (laboratory, computer simulation) or an online setting. We compare a CFA-based policy (interval estimation) to a DLA-based policy (the knowledge gradient).

Predecessor modules: Universal modeling framework
Technical level: Moderate to advanced
Under development

SDA video module thumbnail

Deterministic optimization as a sequential decision problem

Summary: There are many classical deterministic optimization problems that, in practice, are solved repeatedly over time. How a deterministic optimization models are used is often overlooked, especially when they are hard integer or nonlinear programs.

Predecessor modules: Universal modeling framework
Technical level: Requires knowledge of deterministic optimization
Under development

SDA video module thumbnail

Modeling matters

Summary: This module illustrates a published example of a paper that makes the mistake of assuming that you can use information that does not arrive until the future. This type of mistake is not unusual in engineering communities.

Predecessor modules: DLA I
Technical level: Modest
Under development