Schedule

30 lessons of 3 hours, from 23/09 to 30/11/2026. The lessons already given are listed as they actually went; what follows is a plan, and topics keep moving between lessons as the course goes on. What a lesson actually covered is always written in the lesson diary.

23/0928/1011/1130/11
# Date Part Topic
1 Wed 23/09 Intro MPC Course overview, logistics and exam; receding horizon, nominal MPC and its properties, pseudo-code and notation
2 Thu 24/09 MPC MPC in practice: centralized and decentralized, the two time indices, first example, state space models
3 Mon 28/09 MPC The general MPC problem; infinite-horizon constrained control, LQR and its optimal cost
4 Wed 30/09 MPC Why MPC is optimal: dual mode control and the value function
5 Thu 01/10 MPC Conclusions on optimality; first case study: the EV charging service area
6 Mon 05/10 MPC Service area: final formulation and Julia code; losses and indicator variables
7 Wed 07/10 MPC End of the service area example; recursive feasibility
8 Thu 08/10 MPC Proving recursive feasibility; stability, first definitions
9 Mon 12/10 MPC Stability of MPC: Lyapunov theory
10 Wed 14/10 MPC Stability of MPC: the optimal cost as Lyapunov function
11 Thu 15/10 MPC The role of the MPC ingredients; economic MPC; second case study: the building
12 Mon 19/10 MPC Building: shiftable loads, the electricity bill, battery ageing
13 Wed 21/10 MPC End of the building case study; tuning and the Pareto frontier; large-scale EV charging
14 Thu 22/10 MPC End of the large-scale charging example; MPC for traffic control
15 Mon 26/10 MPC Task scheduling in an assembly line; logical constraints at work
16 Wed 28/10 RL From MPC to dynamic programming; introduction to RL
17 Thu 29/10 RL MDPs, value functions, Bellman equations
18 Mon 02/11 RL Dynamic programming: policy and value iteration; numerical experiment
19 Wed 04/11 RL Monte Carlo and temporal-difference methods; Q-learning
20 Thu 05/11 RL Function approximation and deep RL (DDPG)
21 Mon 09/11 RL Hybrid MPC–RL algorithms; learning the weights and the terminal cost of an MPC
22 Wed 11/11 ID Introduction to identification; statistical estimators and their properties
23 Thu 12/11 ID Linear and weighted least squares; regularization
24 Mon 16/11 ID Maximum likelihood, Bayesian estimation, Cramér–Rao bound, nonlinear least squares
25 Wed 18/11 ID Recursive least squares; dynamic models (ARX, output error), predictors
26 Thu 19/11 ID Prediction error methods; experiment design; model validation
27 Mon 23/11 ID Lab: full identification workflow on real data
28 Wed 25/11 ID Kalman filter, EKF, moving horizon estimation
29 Thu 26/11 ID Data-driven predictive control (Willems’ lemma, DeePC); time series and advanced methods
30 Mon 30/11 Wrap-up Course recap, exam information, identify-then-control demo