Course Title: Decision Making Under Uncertainty

95-760 (Heinz) — Spring 2025, Mini 3 (6 units)

Intro to modeling and computational methods for decision support. The first half covers deterministic optimization (LP, network flows, integer programming). The second half introduces risk and uncertainty (probability, simulation, stochastic programming). Emphasis on modeling, Excel-based computation, and analytical interpretation.

Expected Course Schedule (Spring 2025)

Week	Lectures / Activities	Readings	Deadlines / Exams
1	Lecture 1: Introduction to Operations Research; Linear Programming Lecture 2: Linear Programming	Optional: OR — A Catalyst for Engineering Grand Challenges Ragsdale Ch. 2; Ch. 3 (3.0–3.5, 3.7–3.14)	HW1 out — Linear Programming & Sensitivity Analysis (Thu Wk1)
2	Lecture 1: Sensitivity Analysis Lecture 2: Network Flows	Ragsdale Ch. 4 (4.0–4.6); Ch. 5 (5.0–5.7)	HW1 due; HW2 out — Network Flows
3	Lecture 1: Integer Programming (1/2) Lecture 2: Integer Programming (2/2)	Ragsdale Ch. 6	HW2 due
4	Lecture 1: TBD Lecture 2: Exam 1 (both sections)	Bertsimas & Freund, Data Models and Decisions (pages on Canvas)	HW3 out — Integer Programming
5	Lecture 1: Probability Modeling Lecture 2: Sampling and Simulation (1/2)	Ragsdale Ch. 12	HW3 due
6	Lecture 1: Sampling and Simulation (2/2) Lecture 2: Stochastic Programming		HW4 out — Simulation and Optimization
7	Lecture 1: TBD (Optional) Lecture 2: Exam 2	Ragsdale Ch. 14	HW4 due

 

Course Title: From Data to Action

94-867 (Heinz) / 19-867 (EPP) / 12-768 (CEE)

This is a 7-week intensive course on designing full-stack decision systems that prioritize fairness, adaptability, robustness, and human impact. We explore end-to-end pipelines where data fuels actions. Core themes include civic planning, public health operations, and mobility system optimization—universal basic mobility, routing overrides, and last-mile service design.

The course emphasizes implementation: not just what to predict, but how to act. Students build dynamic decision systems and simulate how these systems respond to real-world uncertainty, from emergency hospital staffing to adaptive AI interfaces to route planning under equity constraints.

Prerequisites: Prior coursework in Optimization/Management Science, intermediate Python, and basic ML.

Weekly Plan (Fall 2025)

Week	Date	Topic / Activity	Techniques & Tools
From Human to Machine: Modern Tools for Data → Model → Action → Impact
1	M 8/25	Course Overview; LP & MIP Review	Gurobi
	W 8/27	Gurobi (Cont.); SIR	Optimization API; Compartmental models
	F 8/29	Recitation: Case Code Walkthrough	Hands-on implementation
2	M 9/1	Labor Day — No Class
	W 9/3	Exploration vs Exploitation — Multi-Armed Bandit	Bandit algorithms
	F 9/5	Recitation: Case Code Walkthrough	Hands-on implementation
3	M 9/8	Reinforcement Learning & Dynamic Programming — Backward Induction	DP; policy evaluation
	W 9/10	Q-Learning	Model-free RL
	F 9/12	Recitation: Case Code Walkthrough	Hands-on implementation
From Machine to Human: Tuning Tools to Capture Implicit, Dynamic Human Considerations
4	M 9/15	Machine Learning for Decision Making — Predict-then-Optimize	Integrated prediction + optimization
	W 9/17	Deep Learning & Smart Predict-Optimize	Deep architectures; SPO
	F 9/19	Recitation: Case Code Walkthrough	Hands-on implementation
5	M 9/22	Generative AI for Decision Making — Project Proposal Feedback	Project scoping
	W 9/24	GenAI for Decision Formulation & What-If Analysis	Prompting; planning; sensitivity
	F 9/26	Mini Lecture: GenAI Pretraining and Finetuning	LLM pretrain/finetune concepts
6	M 9/29	Inverse Decision Making — Human Preference Alignment & Inverse Optimization	Inverse optimization; alignment
	W 10/1	Course Review & Synthesis	Integration
	F 10/3	In-class Quiz	Assessment
From Data to Action — Final Project
7	M 10/6	Project Presentation	Capstone
	W 10/8	Project Presentation	Capstone