Madrid's new plan changes what counts as an argument towards interactions in the Built Environment. The PEM sets measurable objectives instead of prescribing means, checks them through indicators, and will verify compliance in a city-scale simulator from the design phase. Whoever wants to build will have to prove performance; whoever reviews will have to validate the proof. F05 trains both sides of that exchange. Students work on six topics taken directly from the plan's strategic lines: photovoltaic potential, heat island and climate refuge, noise, pedestrian proximity, energy retrofit, structural feasibility; each taught at two fidelities on the same problem: the validated engine first, then the fast surrogate, then the comparison that quantifies the error. Every result is framed before/after against the existing city; a raw number out of context means nothing.
AI Simulation
Structural, environmental, and multi-system simulation — with AI validation
The module
Gallery
Objectives
- Run the validated engine for a PEM topic and know what and how it computes.
- Run the fast surrogate of the same problem and state its error first
- Frame every result before/after against the existing city.
- Sensitivity analysis: turn exploration into evidence
What you will be able to do
Learning outcomes
The student runs one PEM topic at two fidelities and quantifies the gap between them: the validated engine against the fast surrogate, error stated before either is used.
Professional outcomes
Madrid's PEM will process planning applications through indicators, with automated checks from the design phase. The practitioner who can produce that evidence, and validate a consultant's, stops buying analysis, wasting feedback rounds and starts arguing with it. That skill is directly billable: performance narratives for tenders, sun-rights and comfort evidence for planning submissions, retrofit cases built on the city's own numbers, schematic structural sanity checks before the engineer is appointed. Rotterdam's performance-based high-rise rules produced a generation of computational specialists; Madrid is about to run the same experiment at full-city scale. The module gives an independent practitioner a credible, defensible skills on the plan their own city will be governed by.
Methodology
Learn by doing, on the student's own real project. No software tutorials: a pre-built starting package is provided and base resources are given apart. Sessions combine a short input, guided practice and critique.
Concepts
Session by session
- What is simulation in the new Madrid
- The PEM, the performance-based regime, the six municipal topics, the brief. Masterclass: Leo Stuckardt (MVRDV) on Rotterdam's densification policies and the development of SolarScape and RoofScapes
- First family of analysis: ray-collision
- Anatomy of a ray caster: intersection tests, blocking rays, sky subdivision. First validated radiation run against a surrogated model (Cyclops). Once engine is understood, then students vibe-code their own checker and bound its error against it (Privacy, view quality, etc.)
- Fast vs. Accurate
- The pre-run microclimate engine (EnviMET) against its live surrogate (infrared.city), error on the table. A space-filling batch opens the design space; sensitivity ranks its drivers; the optimizer searches it. Week-1 lock: one simulation, one exploration, a signed sensitivity note. Hackathon brief handed out.
- Make your own surrogate
- Understand when a network can replace a solver, and how it fails: silently. Implement a ready to run ANN test case from Computational Intelligence Course of TU Delft
- The remaining engines, and the consultant drill
- Pedestrian flows, noise, energy retrofit, structural feasibility at exposure level. Then the drill: re-run a consultant's report and argue with its data (pixels).
- Hackathon: feasibility under the PEM
- Using a pro-forma proxy, run 6 simulations that prove max development yields within performance bounds in an existing prlot in madrid. Jury with 5-min per team to defend their case.
- Guest lectures, when scheduled, are additional to these sessions. The coordinator introduces the guest, who gives a talk, followed by debate and work with students. For this edition, invitations are directed at practices where simulation is embedded in design decision-making at scale — computational teams at global engineering consultancies and applied research groups inside major architecture offices — so that students see, on real built work, how the validation discipline taught in the module operates under professional stakes, and can submit their own pipelines to that same scrutiny.
Deliverable
Faculty and resources
Module Faculty
Learn this in MIAWS
This module is part of the 18-week postgraduate program. Work with your own real project from week 1. First cohort February 2027.
Talk to MIAWS AI →