These kinds of tools are mostly about the input constraints, heuristics, and weights. I'd love to see more about this with more practical constraints (rectilinear rooms and hallways, minimum area for cafetaria/gym, etc) and additional heuristics, such as window area in classrooms, number of hallways/intersections for ease-of-navigation, and building costs beyond just material use.
Yep. The "optimized" floorplans would be an order of magnitude more expensive to build than the original. I mean, drywall alone would take a lot more time with all the cuts, mudding and sanding more joints. The thought of framing out these walls makes my head hurt.
And even if you had a building process that made circular/n-gon-based rooms practical, furnishing and laying out those rooms would be awkward. See the problems people have furnishing dome houses.
I'd least envy the people trying to arrange shelving in the library...
Still, interesting. I'd be curious to see results as he adds more practical constrains.
Dome home: most rooms will have 2 or 3 flat walls with 1 or 2 right-angle corners, so put rectangular furniture there.
Classrooms and library at the proposed school: the size of the room dwarfs the size of the furniture, so the problem is much like complaining that the letter "O" can not be made from square pixels. Somehow, we get our letter "O".
(I'm not even sure if it's real or feasible; I just remember a bunch of pop news articles with a video of a large CNC machine depositing concrete to construct a building.)
There's a YouTube channel, "My Little Homestead" where a family builds various house-additions and structures out of stucco-covered dirt-filled sandbags.
Dunno how that sort of construction technique pans out long-term or large-scale but you could get the sort of floorplans the GA generated out of it.
You have to build formwork to construct with in-situ concrete, so there's still carpentry (or equivalent) involved.
It's a good point that existing architectural concrete 3d printing creates "mass concrete" not reinforced concrete, but mass concrete can be structural.
Rebar makes earthquakes worse, as seen in California's freeway collapse. The rebar depends on the concrete to prevent corrosion, but that is only good for roughly 50 years. Rust causes expansion, which cases cracking. The outer concrete falls away, leaving rebar to get crushed under the load. The obvious fixes, like stainless steel rebar, have thermal expansion coefficients that don't match concrete, so instead you get cracking even without corrosion.
Lots of Roman stuff is still standing in areas that get earthquakes. The solution is to use a conservative design, with arches and thick walls. Domes are good. We can improve on this with 3D printing, using a structure like mammalian bone: solid near the surface, and spongy in the middle.
Would you be able to provide links to modern structures built with the methods you're describing or research about them? Does this conservative design resemble something like Gaudi's Sagrada Familia, or is that the wrong way to think about it?
Arch dams might be the only type of "contemporary" looking structure that is habitually made in this way out of unreinforced concrete.
Although Gaudi was interested in structural optimization (using catenary models), he is an outlier in terms of design. He didn't comprehensively consider seismic aspects, though apparently he didn't do too badly: https://blog.sagradafamilia.org/en/divulgation/seismic-activ...
"Primitive" structures and other structures engineered to avoid tension in the concrete. Concrete doesn't have zero tensile strength. Apparently it's a few MPa, so a fifth or a tenth of the compressive strength. http://www.engineeringtoolbox.com/amp/concrete-properties-d_...
The large dome of the (Roman) Pantheon is unreinforced concrete. Most compressive structures built before the modern period (arches, domes, buttresses) were not reinforced with tension elements, just plain masonry.
These guys are building small houses with a thing that prints in layers just like a home 3D printer, using mortar instead of plastic. https://www.youtube.com/watch?v=wCzS2FZoB-I It looks like there's manually placed metal mesh for reinforcement.
Dunno what the quality's like, but it's there. It does seem like it would be tricky to scale to structures larger than the practical footprint of their printer, but maybe the thing could be mounted on wheels.
You could probably print in sections as well. This looks like it would be good for single story structures and possibly foundations/footings for larger 2 and 3 story structures.
But would it still cost more if the building were made of prefab-stamped “IKEA pieces”, rather than stick-built? I.e., if you had to build the house 100 times (for a townhouse development, say), and so “factored out” the assembly of the framing and drywall and etc. for each curve and corner, would the labor costs then approach equivalence?
No, because those joints still need to be created somewhere in the process. It might be cheaper than the one-off build, but it wouldn't be cheaper compared to a more efficient design "factored out."
These types of exercises are useful not to find their solutions, but to reveal the shortcomings of the model used to create them - and our thus our own mental models!
Building design is older than humankind, and there's tens of millenia worth of reasons why they are the way they are. But those reasons aren't always clear. Essentially modern floor plan design is the output of a black-box machine (human) learning algorithm.
The flawed output here, and the comments, help us point out what some of the missing variables are: egress, airflow, construction cost, etc. Do enough of this and you can get a vastly improved model.
> The flawed output here, and the comments, help us point out what some of the missing variables are: egress, airflow, construction cost, etc.
...or maybe illustrate that some of the existing external constraints are stifling architecture as a useful art form?
Modern structures have to be fully designed before they are built, whereas most traditional structures were designed and built at the same time, and the design could evolve as needed to fit the environment. Now, we do everything with straight lines because it's easier to make plans and estimate materials and communicate with builders and file for permits and verify code compliance and so on, but you can do much more interesting and complicated designs if you don't have to communicate the design with a human at every step.
I think sometime in the not-too-distant future we'll have practical machines that can construct buildings designed by software to conform to the features and limitations of the building site and the desires of the future owners. If there's no need to communicate the design to a human other than "does this rendering look good?" then we don't need right angles everywhere just to make life easy for the draftsmen and carpenters.
This seems kind of related to something Christopher Alexander said about twenty years ago (which was linked from HN recently [1]), that current design and construction methods and business models have basically made good architecture well-nigh impossible to achieve on a wide scale. I don't know what he would think of using his architectural pattern language as a set of algorithm heuristics, but it's one possible way forward.
I disagree: the issue isn’t with the expressivity of the model, but with the objective problem and constraints set by the author. You can quickly adjust the algorithm to product floor plans resembling modern buildings by penalizing construction materials.
Many of the social insects -- ants and termites and bees, especially -- build "buildings" which have heterogenous rooms specialized for different purposes.
> there's tens of millenia worth of reasons why they are the way they are.
Eh, not quite: all of these "reasons" are adaptations to a specific condition. Many of which only exist since recently (for example, reinforced concrete enforces certain adaptations).
Evolution is not working towards one static best solution. It is always a process of adaptation of an existing solution to the current context.
I don't disagree. However no one civilization would have been able to invent, for example, a skyscraper of reinforced concrete, in a vacuum, without everything that came before that.
Sure, it looks cool. But there are so many other properties of a home that aren't accounted for in this model (mostly aesthetic ones, like cozyness, feelings of safety/comfort, how the plan interacts with light, etc). One day we'll live in homes with automated floorplans, I'm sure of it. But it won't happen (or at least, we won't be happy about living in such things) until the day we figure out ways to deal with more dimensions than efficiency (who knows, maybe crowdsourcing ratings of CGI rendered apartments will be the answer, somehow it'll be solved).
I was at a vehicle AI conference early this month discussing how AI will only output based on the quality of its inputs. And the fear coming out of that conference was if space wasn't correctly dedicated or allocated, the AI will optimize the systems as much as it can to squeeze out every single ounce of efficiency.
Qualitative items such as the ones listed are important for customer focused environments, however, I'm not sure if AI can account for such factors.
> Qualitative items such as the ones listed are important for customer focused environments, however, I'm not sure if AI can account for such factors.
One of the nice things you can do with optimization problems is plug humans into the loop as oracles. Often, 'we know it when we see it', and we can do pairwise comparisons of 2 possibilities. So you can train a ML model based on win/loss comparisons and it'll learn to take into account the softer qualitative aspects via preference learning.
But also "Deep TAMER: Interactive Agent Shaping in High-Dimensional State Spaces" https://arxiv.org/abs/1709.10163 , Warnell et al 2017.
You can even pair it with EEG or brain scans for implicit ranking: "Towards personalized human AI interaction - adapting the behavior of AI agents using neural signatures of subjective interest" https://arxiv.org/abs/1709.04574 , Shih et al 2017.
Like I mentioned, I have high hopes modelling vaguer aspects of customer preference can be incorporated by training a model on rendered generative design (e.g. with data from amazon turk), which can then be used as maybe some kind of penalization on the efficiency loss function, multiplied by a weight to give us a choice in the efficiency-comfort tradeoff-space.
There are solutions that I think show potential out there. I don't think our future AI designed world necessarily need to ignore difficult-to-quanitify dimensions like aesthetics. (Though amazon turk is expensive, especially in developer man-hours, so I can understand if that won't always be done.)
It all comes down to how explicitly preferences are originally understood and then if the reward function can incorporate implicit analysis.
There have been recent studies about AI powered shirt design - the original input uses existing designs in terms of color and shape rather than the basic naive description of requirements that an engineer would give. Then the designs can be assessed by a review board or put up on a site and not produced until some n quantity of purchases.
You wouldn't try to detect cats in images without labelled data why would you try something MUCH harder without labelled data?!?!?!?!?!
Why are you sure that at some point in the future, we will live in homes with AI-generated floorplans?
It's a bit like saying that poetry will mostly be written by computers at some point in the future, or humans will mostly have sex with robots in the future. There's going to be more to this than whether the aesthetic experience is acceptable.
The author describes the results as 'irrational,' but I'd hardly agree with the negative connotation of that word: I think that the generated floorplans look lovely, and I wish that I'd gone to school somewhere like that. Far from irrational, I think it's quite rational to try to design buildings to be pleasant to live in, not just easy to draw with a straightedge.
Also, the generated rooms look much like the sort of rooms one would build with cob. A cob schoolhouse could be awesome.
You need good sightlines for the kids, room for desks, and need to ensure kids aren't isolated in a corner. Plus light (not just to a small courtyard), ability to escape in case of fire or other emergency!!!!!!! and for minimal noise intrusion from other classrooms.
Minimal space and building material are just not the right constraints and speak to absolute ignorance of the actual problem on the part of the designer.
it seems like it speaks more to the disinterest in providing an actual solution. The article is providing a possible route to finding a solution; not giving it. I mean he literally only optimizes on three parameters: windows, escape routes, and traffic usage.
He even clearly states his goals: The creative goal is to approach floor plan design solely from the perspective of optimization and without regard for convention, constructability, etc. The research goal is to see how a combination of explicit, implicit and emergent methods allow floor plans of high complexity to evolve.
Nowhere does he claim this is an implementable solution, or meant to be one... just an interesting one, given a set of priorities.
I will have to disagree with the author’s characterization of these floor plans as irrational.
As pointed out by commentator from the r/ML subreddit, the plans resemble biological arteries — which intuitively make sense given that the objective was to minimize distance to nearest fire exit. The analog is that arteries try to maximize surface area to volume ratio to ensure blood flood.
I thought classrooms required an emergency exit. The original layout is probably specifically designed to give every room an exterior wall. Still, very interesting project.
The original layout looks quite reasonable. It seems to optimize giving classrooms many windows for natural light and fresh air. Additionally the layout mostly keeps noisy areas such as the gym and cafe away from classes.
I'd like to apply ML to individual and group behavior in a building (biometrics, activity, location, laughter, concentration) and create a built environment that satisfies unconscious preferences for both static and dynamic uses. The best possible terrarium for humans.
'cinderblock utilitarianism' is exactly in-line with _The Timeless Way of Building_. Four square walls and a peaked roof? Job done for the next 300 years.
Mea culpa. I think I was using “The Timeless Way of Building” as shorthand for “Pattern Language”. Or at least, I was thinking of the list of desirable properties rather than the method of construction.
Of course, if you designed and built such a thing all at once using a giant 3D printer, I guess you'd be building in an entirely new way, which would be completely different :-)
Did you miss the part about "light on two sides of every room"? These evolved designs seem optimised for nothing that actually matters to human comfort.
I took this as an initial exploration of genetic algorithms for architecture. As such, the idea is more interesting than the specific parameters and weights chosen here.
You'll notice as you scroll down that this article _does_ vary the parameters and weights. For instance, the one that optimizes for windows tends to create interior courtyards.
I could imagine adding parameters and weights for thresholds, areas for crafts, light on two sides of rooms, rectilinearity, water shedding, etc. etc. etc.
But I love the idea that what comes out the other end is wildly weird and yet functional, and feels organic and nest-like rather than square.
I posted my comment after reading the article, which included optimized-for-windows results. Of course, others may have had it sitting around in a browser tab for a while…
I've been in a few apartments that had novelty floor plans without right angles. All of them were terrible. Granted I have not been in one owned by somebody rich or skilled enough to only have custom made furniture.
I’d be curious to see the effect on airflow, though. I’d gladly pay for custom furniture if the design of the house meant never sitting in a cloud of my own CO2 (without any need for fored-air ventilation like in airplanes/submarines/bunkers.)
I used to go to the IBM Somers campus semi-regularly (IM Pei pyramid-shaped buildings essentially.) It was filled with these weirdly shaped conference and other rooms that made pretty bad use of space.
The Hogwarts floorplan improves on this, reorganizing in real time to the needs of individual students and faculty. Any sufficiently advanced set of genetic algorithms, sensors and moving partitions is indistinguishable from magic.
i am a product of american public school. in every school all the classrooms (except portables) had a wall that was mostly windows. you can see thats also true in the original school design. i’d rather have a ton of windows than shorter distance to walk. you could feel the enclosed difference in portable rooms
His genetic examples look like European cities compared to the actual floor layouts, which are more like New World ones designed on a grid layout by central planners.
Perhaps there's wisdom in allowing cities to grow organically/chaotically after all.
There is, it has been proven that on average, if you select two points in European cities the average distance by road, number of junctions etc is smaller than two randomly selected points in a grid system.
It seems like a grid system is superior if you have no knowledge of the city, but a "organic" european layout is functionally superior assuming people always choose perfect routes.
On a related note I’d love a way to optimize kitchen layout. What items to store in cabinets and drawers to minimize dishwasher unloading time, optimize cooking.
I can see this kind of architecture being the next fad in architecture as a kind of next step from the Gehry-style architecture. I can totally see someone pr firm or tour guide bragging "our building was designed by a computer algorithm to be optimized for our needs..."
"...on the day it was designed. Of course, by the time construction was completed there had been 2 re-orgs and a pivot so now we're just making the best of things"
They give the impression of the building being alive. And therefore your presence is potentially a threat that it will deal with and you'll be trapped for an eternity in an endless maze of reconfiguring hallways and dead ends.
Well I wasn't feeling this way but now I am... Seriously though I prefer the more organic feel. I don't really like square hallways and I wouldn't mind being in a rounder room. though it would need to take into the actual requirements of the rooms. A shop class needs both a lot of space and square enough walls/floor plan that they can fit in equipment without a ton of wasted space.
Weird, I had the opposite feeling, it looked comfortable and organic and natural to me. Maybe it depends on how regularly you get out into and appreciate nature.
This is great. I did a major project in eighth grade about designing a school to make traffic flow more logical and the strict times they specified for student travel achievable. We were expected to leave a class, travel to our lockers, accomplish any necessary bathroom tasks and to travel to the next class in four minutes and 20 seconds. Failing to do so was a disciplinary matter, but it did not appear that the school or our schedules were designed with this sort of urgency in mind.
The main focus of my solution was to make the school a torus. Like many of my youth projects (writing a text adventure game...), I did not have the knowledge to solve the problems scientifically.
I'm an architect in germany, working as project-lead for the initial stages of medium sized buildings till building permit and therefore responsible for the conception and its validation in all terms of structure, fire-protection, installations, etc.
I'd like to make my point that any kind of over-optimization of one or more aspects will put one of the biggest virtues of truly good and functional architecture at risk: being adaptive for a wide variety of future changes (use, program, technology, climate, energy resources, partitioning, shrinking, expanding, etc.).
The actual task of designing a building is to find the right balance in a myriad of parameters, which sometimes create synergies (think sunlight and heating), but often are just one step away from undesirable impacts (think sunlight and overheating). Flexibility in architecture always results out of generous tolerances and robustness - which is always in danger to be eliminated by optimization for limited scenarios.
Also many aspects of the design logically derive from each other: if I plan a school with natural ventilation, it'd be probably a good idea to have windows in two sides of the room, that can exchange the whole air of the room in a 5min break. If I opt for a mechanical ventilation this advantage would be gone and the disadvantages of not having a more compact cubature would override and result in a completely different layout.
While I appreciate all kind of tools that give me an insight into complex interdependencies (how do floorplans with optimized A,B,C look like?) i think that good architectural solutions need humans to make a tailor-made decision based on a bigger picture of our society that has the chance to be valid for some decades(centuries?). Good architects choose to rely on typologies that evolved from history for this difficult task and transform them when needed.
I'd be curious if the approach of OP could be used backwards as a software based analysis what details make successful typologies actually successful.
On a side note: It's pretty interesting that the resulting floorplans of OP are somewhat similar to the traditional arabic city structure (google traditional damascus city center and zoom into the still intact quartiers).
I work in 3D drafting & design, and CNC programming. In my life I've seen woodwork go from almost always boxes, to heavily jellybean, organic shapes. Because we have the tools to do it, people want it. Any less looks dated.
Houses will go the same way as the tools become available. In the yacht business a new design took a room full of men a couple years to design and draw, so they were slow, serious business. Now it takes three guys at computers six months. So people expect it: a new design every year at the boat show. Go two years without and bad rumors circulate.
Almost 2 decades ago I wrote a perl script to automatically optimize open space floor plan using genetic algorithm.
The idea was to have a function, for each office worker in the office that would grade the place they are sitting in (distance to toilet, printer, lighting, space around desk, etc.) and have algorithm evolve open space plans to have best overall satisfaction and also make sure that there are no places that have very low score.
Now, this algorithm could only take existing floor plan, it could only place furniture on existing floor with existing walls but it was still nice excercise.
There's a whole academic sub-discipline in Mechanical Engineering devoted to optimizing geometries like this. Shape Optimization tries to find shapes with particular properties, e.g. minimizing material use while maximizing shear strength. You'll find hundreds of papers going back decades if you want to dig into it:
You’d need to carefully study the safety implications. People already get disoriented in fires when they have to move in a straight line. Add winding and branches and it could cause some issues.
The alternative to a cultural standard is custom pieces. In this case, the cultural standard is right angles. It could have been 60 degree angles and we would all think hexagonal rooms are ideal.
Chairs, tables, desks, bookcases, shelving systems, and all the things that go on top of them. Paper. Boxes. Books.
I like it, despite the infeasibility. Here are a few more things missing from the plan: 1) The main door should be by the admin offices, for security reasons. 2) The gym has to be rectangular for the basketball court, or large enough that it can contain it. 3) One of the gym's looks slightly non-convex, which is pointless. (Although that might be an artifact of rendering the door.)
>U.S. fire departments responded to an estimated average of 4,980 structure fires in educational properties in
2011–2015, annually. These fires caused annual averages of one civilian death, 70 civilian injuries, and $70 million
in direct property damage.
Nearly seven in ten of the fires in educational properties (69%) occurred in nursery, elementary, middle, or high
So roughly 25 deaths and 1750 injuries in the last 25 years in the US. I think this is amazingly low considering the 4,980 structure fires in schools per year.
But your point is well taken; there's no particularly great reason to, unless your school is so big and your administration is so goofy that kids are constantly running late. That constraint would be better suited to applying these algorithms to offices, hospitals, military bases, prisons, and maybe shopping malls.
The high school I went to (Alvin H.S. in Alvin, TX) was laid out terribly. It was almost 70 acres, and there were dozens of buildings with classes in them.
We were given 7 minutes between classes, and my freshman year I had to cross the campus multiple times per day (more than 2000 feet each way). That meant pretty much running.
On days when it was raining, and you couldn't cut through the grass, or if you were handicapped and had to stick to the 5 foot wide covered paths, you were guaranteed to be tardy and written up or sent to your office (potentially back across campus).
In the nearly 15 years since I left that school, they have consolidated most of their buildings, and only have about 10.
They could definitely have used some sort of planning algorithm to come up with the optimum scheduling for their students.
Right optimized one looks best, because it's the only one that doesn't squirrel any of the big rooms (library, gym, cafeteria) away at the ends of corridors.
I think it's probably a balancing act between placing the big room at the end of a hallway (making that hallway a nightmare prior to any event) and placing it in the middle of the building (making that intersection a nightmare prior to any event). I'm sure it depends largely on the specifics of the room and how often it will be used of course. Less frequently utilized is probably better off further away from the center, more frequently better off closer but with multiple entrances & exits to ease congestion.
