Merge branch 'main' into kwu

Author: kevinwoots

.github/workflows/documentation.yml

@@ -17,7 +17,7 @@ jobs:
       - uses: actions/checkout@v4
       - uses: julia-actions/setup-julia@v2
         with:
-          version: '1'
+          version: '1.11'
       - uses: julia-actions/cache@v2
       - name: Install dependencies
         run: julia --project=docs/ -e 'using Pkg; Pkg.instantiate()'

README.md

@@ -11,6 +11,9 @@
 ## Overview
 This student-led course explores modern techniques for controlling — and learning to control — dynamical systems. Topics range from classical optimal control and numerical optimization to reinforcement learning, PDE-constrained optimization (finite-element methods, Neural DiffEq, PINNs, neural operators), and GPU-accelerated workflows.
 
+## Objective
+Create an online book at the end using the materials from all lectures.
+
 ## Prerequisites
 * Solid linear-algebra background  
 * Programming experience in Julia, Python, *or* MATLAB  
@@ -19,9 +22,22 @@ This student-led course explores modern techniques for controlling — and learn
 ## Grading
 | Component | Weight |
 |-----------|--------|
-| Participation & paper critiques | **25 %** |
-| In-class presentations | **50 %** |
-| Projects | **25 %** |
+| Participation | **25 %** |
+| In-class Presentations and Chapter | **50 %** |
+| Projects (Liaison work & Scribe & Admin & ...) | **25 %** |
+
+**Class material is due one week before the lecture!** No exceptions apart from the first 2 lectures.
+
+**Issues outlining references that will be used for lecture preparation are due at the end of the 3rd week (10/05/2025)!** 
+20 minutes of research should give you an initial idea of what you need to read.
+
+🎯🚲 **Guessing Game**
+
+Here’s how the presentation grading works: we already know the lecture content we expect from you. Any deviations will be penalized **exponentially**. Your mission is twofold:  
+1. **Check your understanding** — use [discussions](https://github.com/LearningToOptimize/LearningToControlClass/discussions) from previous lectures to ensure you’ve mastered earlier topics. We expect lectures to be extremely linked to each other.
+2. **Test your hypotheses** — validate your lecture content by raising and resolving issues, focusing primarily on your *main task issue* (see this example from [class 03](https://github.com/LearningToOptimize/LearningToControlClass/issues/18)).  
+
+All interactions will happen **only through GitHub** — no in-person hints will be given. 
 
 ## Weekly Schedule (Fall 2025 – Fridays 2 p.m. ET)
 
@@ -30,16 +46,16 @@ This student-led course explores modern techniques for controlling — and learn
 | #  | Date (MM/DD) | Format / Presenter | Topic & Learning Goals | Prep / Key Resources |
 |----|--------------|--------------------|------------------------|----------------------|
 | 1  | 08/22/2025   | Lecture — Andrew Rosemberg | Course map; why PDE-constrained **optimization**; tooling overview; stability & state-space dynamics; Lyapunov; discretization issues | [📚](https://learningtooptimize.github.io/LearningToControlClass/dev/class01/class01/) |
-| 2  | 08/29/2025   | Lecture - Arnaud Deza | Numerical **optimization** for control (grad/SQP/QP); ALM vs. interior-point vs. penalty methods | |
+| 2  | 08/29/2025   | Lecture - Arnaud Deza | Numerical **optimization** for control (grad/SQP/QP); ALM vs. interior-point vs. penalty methods | [📚](https://learningtooptimize.github.io/LearningToControlClass/dev/class02/overview/) |
 | 3  | 09/05/2025   | Lecture - Zaowei Dai | Pontryagin’s Maximum Principle; shooting & multiple shooting; LQR, Riccati, QP viewpoint (finite / infinite horizon) | |
 | 4  | 09/12/2025   | **External seminar 1** - Joaquim Dias Garcia| Dynamic Programming & Model-Predictive Control | |
-| 5  | 09/19/2025   | Lecture - Guancheng "Ivan" Qiu | **Nonlinear** trajectory **optimization**; collocation; implicit integration | |
+| 5  | 09/19/2025   | Lecture - Guancheng "Ivan" Qiu | **Nonlinear** trajectory **optimization**; collocation; implicit integration | [📚](https://learningtooptimize.github.io/LearningToControlClass/dev/class05/class05/) |
 | 6  | 09/26/2025   | **External seminar 2** - Henrique Ferrolho | Trajectory **optimization** on robots in Julia Robotics | |
 | 7  | 10/03/2025   | Lecture - Jouke van Westrenen | Stochastic optimal control, Linear Quadratic Gaussian (LQG), Kalman filtering, robust control under uncertainty, unscented optimal control; | |
-| 8  | 10/10/2025   | **External seminar 3** TBD (speaker to be confirmed) | Topology **optimization** | |
+| 8  | 10/10/2025   | Lecture - Kevin Wu | Distributed optimal control & multi-agent coordination; Consensus, distributed MPC, and optimization over graphs (ADMM) ||
 | 9  | 10/17/2025   | **External seminar 4** — François Pacaud | GPU-accelerated optimal control | |
-|10  | 10/24/2025   | Lecture - Michael Klamkin | Physics-Informed Neural Networks (PINNs): formulation & pitfalls | |
-|11  | 10/31/2025   | **External seminar 5** - Chris Rackauckas | Neural Differential Equations: PINNs + classical solvers | |
+|10  | 10/24/2025   | **External seminar 5** - Chris Rackauckas | Neural Differential Equations: classical solvers + adjoint methods | |
+|11  | 10/31/2025   | Lecture - Michael Klamkin | Physics-Informed Neural Networks (PINNs): formulation & pitfalls | |
 |12  | 11/07/2025   | Lecture - Pedro Paulo | Neural operators (FNO, Galerkin Transformer); large-scale surrogates | |
 |13  | 11/14/2025   | **External seminar 6** - Charlelie Laurent | Scalable PINNs / neural operators; CFD & weather applications | |
 |14  | 11/21/2025   | Lecture - TBD | TBD from the pool | |
@@ -51,14 +67,12 @@ Students must provide materials equivalent to those used in an in-person session
 
 | # | Format / Presenter | Topic & Learning Goals | Prep / Key Resources |
 |---|--------------------|------------------------|----------------------|
-| 15 | Lecture - TBD | Quaternions, Lie groups, and Lie algebras; attitude control; LQR with Attitude, Quadrotors; | |
-| 16 | Lecture - TBD | Stochastic optimal control, Linear Quadratic Gaussian (LQG), Kalman filtering, robust control under uncertainty, unscented optimal control; | |
-| 17 | Lecture - TBD | Trajectory Optimization with Obstacles; Convexification of Non-Convex Constraints; | |
-| 18 | Lecture - Joe Ye | Robust control & min-max DDP (incl. PDE cases); chance constraints; Data-driven control & Model-Based RL-in-the-loop | |
+| 15 | Lecture - Shuaicheng (Allen) Tong | Dynamic Optimal Control of Power Systems; Generators swing equations, Transmission lines electromagnetic transients, dynamic load models, and inverters. | |
+| 16 | Lecture - Joe Ye | Robust control & min-max DDP (incl. PDE cases); chance constraints; Data-driven control & Model-Based RL-in-the-loop | |
+| 17 | Lecture - TBD | Quaternions, Lie groups, and Lie algebras; attitude control; LQR with Attitude, Quadrotors; | |
+| 18 | Lecture - TBD | Trajectory Optimization with Obstacles; Convexification of Non-Convex Constraints; | |
 | 19 | Lecture - TBD | Contact Explict and Contact Implicit; Trajectory Optimization for Hybrid and Composed Systems;  | |
 | 20 | Lecture - TBD | Probabilistic Programming; Bayesian numerical methods; Variational Inference; probabilistic solvers for ODEs/PDEs; Bayesian optimization in control;  | |
-| 21 | Lecture - TBD | Distributed optimal control & multi-agent coordination; Consensus, distributed MPC, and optimization over graphs (ADMM). | |
-| 22 | Lecture - TBD | Dynamic Optimal Control of Power Systems; Generators swing equations, Transmission lines electromagnetic transients, dynamic load models, and inverters. | |
 
 ## Reference Material
 

class01/background_materials/math_basics.html

@@ -39,6 +39,8 @@ md"
 |  Lecturer   | : | Rosemberg, Andrew |
 |  Date   | : | 28 of July, 2025 |
 
+Special thanks to **Guancheng Qiu** for helping fix some of the code!
+
 # Background Math (_Welcome to Pluto!_)
 
 This background material will use Pluto!

class01/background_materials/math_basics.jl

@@ -37,6 +37,8 @@ md"""
 |  Lecturer   | : | Rosemberg, Andrew |
 |  Date   | : | 28 of July, 2025 |
 
+Special thanks to **Guancheng Qiu** for helping fix some of the code!
+
 """
 
 # ╔═╡ eeceb82e-abfb-4502-bcfb-6c9f76a0879d
@@ -436,21 +438,32 @@ begin
  [7  1  3  9  2  4  8  5  6];
  [9  6  1  5  3  7  2  8  4];
  [2  8  7  4  1  9  6  3  5];
- [3  4  5  2  8  6  1  7  9]])
-
-	anss = missing
-    try
-        anss = (
-            x_ss   = haskey(sudoku, :x_s) ? JuMP.value.(sudoku[:x_s]) : missing,
-        )
-    catch
-        anss = missing
+ [3  4  5  2  8  6  1  7  9]],)
+
+    anss = (;
+        x_ss = haskey(sudoku, :x_s) && JuMP.is_solved_and_feasible(sudoku) ? JuMP.value.(sudoku[:x_s]) : missing
+    )
+
+    # Convert 3D binary matrix to 2D solution matrix
+    function convert_3d_to_solution(x_3d)
+        if ismissing(x_3d)
+            return missing
+        end
+        solution = zeros(Int, 9, 9)
+        for i in 1:9, j in 1:9, k in 1:9
+            if x_3d[i, j, k] ≈ 1.0
+                solution[i, j] = k
+            end
+        end
+        return solution
     end
 
-    goods = !ismissing(anss) &&
-           all(isapprox.(anss.x_ss, ground_truth_s.x_ss; atol=1e-3))
+    solution_matrix = ismissing(anss) ? missing : convert_3d_to_solution(anss.x_ss)
+
+    goods = !ismissing(anss) && !ismissing(solution_matrix) &&
+           all(isapprox.(solution_matrix, ground_truth_s.x_ss; atol=1e-3))
 
-    if ismissing(anss)
+    if ismissing(anss.x_ss)
         still_missing()
     elseif goods
         correct()
@@ -578,8 +591,8 @@ begin
 model_nlp = Model(Ipopt.Optimizer)
 
 # Required named variables
-@variable(model_nlp, x)
-@variable(model_nlp, y)
+@variable(model_nlp, x_nlp)
+@variable(model_nlp, y_nlp)
 
 # --- YOUR CODE HERE ---
 
@@ -707,7 +720,7 @@ begin
     # Decide which badge to show
     if ismissing(ansd)               # nothing yet
         still_missing()
-    elseif x == 25.0
+    elseif ansd == 25.0
         correct()
     else
         keep_working()
@@ -721,8 +734,8 @@ begin
 	ans3=missing
     try
         ans3 = (
-            x   = safeval(model_nlp, :x),
-            y   = safeval(model_nlp, :y),
+            x   = safeval(model_nlp, :x_nlp),
+            y   = safeval(model_nlp, :y_nlp),
             obj = objective_value(model_nlp),
         )
     catch
@@ -799,7 +812,7 @@ end
 # ╟─bca712e4-3f1c-467e-9209-e535aed5ab0a
 # ╟─3997d993-0a31-435e-86cd-50242746c305
 # ╠═3f56ec63-1fa6-403c-8d2a-1990382b97ae
-# ╟─0e8ed625-df85-4bd2-8b16-b475a72df566
+# ╠═0e8ed625-df85-4bd2-8b16-b475a72df566
 # ╟─fa5785a1-7274-4524-9e54-895d46e83861
 # ╟─5e3444d0-8333-4f51-9146-d3d9625fe2e9
 # ╠═0e190de3-da60-41e9-9da5-5a0c7fefd1d7