Using Math To Design Cities And Supercomputers

caek writes "If you've played Sim City you've wrestled with one of the problems faced by supercomputer designers. Unfortunately there's no GameFAQs.com for the technical staff at Japan's Earth Simulator or Srinidhi Varadarajan and colleagues at Virginia Tech. True enough, they won't have to deal with rising crime or Godzilla but, as hinted at in a recent paper in Journal of Physics A, the physical layout of a massively parallel supercomputer is fundamentally the same problem as minimizing the time commuters spent stuck in traffic jams. Read the rest of my kuro5hin article for a popular explanation."

Comments

Why not?

[danratherfan]

Just make the city a big parking lot. With roads that lead to nowhere?

Excellent..

[Kiriwas]

The most powerful SimCity bot ever created!! Oh, and I suppose it can be used for other stuff too.

Does this surprise anyone?

[Marxist Hacker 42]

It doesn't surprise me. I've been saying for years that the algorithims used for network and resource optimization have real-world applications; that the reverse is also true is simple logic. Now- if we could only apply *nix resource allocation algorithims to food, clothing, shelter, and medical care; perhaps then we'd actually have an economy that works for the people instead of a people who work for the economy.

Chicken-and-egg problem! ;-)

[PaulBu]

At the time when I was involved with preliminary design of HTMT petaflops supercomputer (yes, it is petaflops, as in million gigaflops, see, for example, here), anyway, one of the problems which would require a supercomputer with this this kind of performance was real-time optimization of car traffic in a city the size of Manhattan, NY.

Paul B.

Re:Chicken-and-egg problem! ;-)

[AmericanInKiev]

So the question then is - how do you layout a city such that 10 million computer experts can all get to one place and design a super computer capable of calculating the design of such a city.

Idea: remove the vehicle subsidy - which is killing trains, buses, and bicycles.

Then you get mass transit.

In the end - the real barrior to optimizing streets will be the arab oil cartells screaming about the loss of oil revenues to keep those compluters stuck in traffic burning up time, space, and O2.

AIK

The Important thing is Rail

[AmericanInKiev]

Cities should layout their High Speed Bus (er Train) routes early

This will provide the contract builders need to build high density housing along high speed corridors - rather than randomly.

The key to getting commuters out of traffic - is only in part - optimizing their route.

The real key is getting them out of their damn cars.

(Electric Bicycle commuter speaking)

AIK

Re:The Important thing is Rail

[merdark]

Yes, that's all nice, but it doesn't help the many many *existing* cities any. They already have buildings sprawled all over the place.

If you are encouraging building more NEW cities, then I have to totally disagree with you. We need to fix the ones we have somehow. You can't get people out of their cars unless you provide them with a humane way of alternate travel. Trains are good, provided they actually reach people (hard), and are not overcrowded like the subways in Toronto. I voluntarily worked an extra hour every day just to avoid rush out on the overcrowded subways. That kind of crappy mass transit just won't cut it.

Re:The Important thing is Rail

[rewt66]

I read somewhere an article where somebody from Chicago was giving advice to Seattle on their new transit system. The author said that the key to getting people out of their cars is rapid transit - defined as "faster than driving". Just having transit isn't enough. People will still take their cars because it's faster, which equates to more convenient. But if the transit is faster than driving, people are more inclined to take it.

San Francisco added a nice touch to this with their BART system. In some places, they build the rapid transit right up the freeway median. When you're stuck in traffic on the freeway, and the train blows past you at 80 MPH, it tends to make you think, "I wish I was on that!"

Re:The Important thing is Rail

[AmericanInKiev]

Mod this Up.

Often without saying it - That MUST be the effect of shared transit.

This is the primary reason that buses suck.

Buses must have HOV lanes to be a considered alternative.

The other is the cost of parking - London has fees just for entering downtown.

Cities should NOT provide subsidised parking.

They should put that money instead into - as you say - Rapid Transit.

AIK

No computer design required, just brains.

[Chemisor]

Look at this city design. Instead of trying to create better routing of commuters, it eliminates the whole problem of commuting.

Re:We need to fix the ones we have somehow.

[merdark]

2. After they come back - they can vote on a central carfree zone in their city.

I don't know about a totally car free zone, trucks still need to be able to make deliveries for instance. But we definatly need more walking streets like in Europe. Driving downtown in most cities is futile anyways, and walking streets add so much character and class to a city. :)

The city should lay down some serious linear transit routes - in order to encourage as much brownfield redevelopment, and density conversions as possible.


Yes, more trains/subways are needed.

Low density neighborhoods in the line of High Density Trasit should be scheduled for redesign "with all deliberate haste".


This wouldn't fly I don't think. You can't just kick people out of their houses, as convenient as that may be.

We should allocate natural beauty for High Density Housing - rather than the other - which is to reward ineffecient design with luxurious vistas.

This is not as simple as you make it out. For instance, here in Toronto we have 'nice' neighbourhoods. Those neighbourhoods are nice precisly *because* they have no high desnsity housing. Low density housing areas can have lots of nice trees and be located in areas with rolling hills. High density housing by it's very nature needs flat areas and tends to ruin nature directly around it. Basically, I'm saying that the low density areas are not nice because they have nice vistas, most don't.

Also, in Toronto high density housing IS being built along the lake front. But there are two problems: there is no mass transit from that area (!) and the high density houseing obscures the view from previous high density housing built just a few lots inland. So it's only possible for a select few high density buildings to have nice views.

a cellphone GPS based service which garentees a small electric car is waiting for you at at the terminal.

A similar service which can pick you up at the terminal and take you and a few others to work - Like a disney golf-train.

Given the cellphone can know where you are - and where you need too get - it could optimize a last mile service based on current demand and requirements - rather than a fixed - use it or lose it schedule.

Nice ideas, but I'm skeptical that the oil/car companies would let this fly. :( People have tried selling this idea, and cities just don't go for it. Maybe it needs some more voices calling for it or something.

Re:We need to fix the ones we have somehow.

[AmericanInKiev]

Thanks for your perspective.

I think you may have glossed over a point.

The City must establish LINEAR High speed transit corridors.

Because only LINEAR transit routes can achieve - Rapid - as in faster than an SUV - transportation (After you figure in the cost of waiting around)

Deliberate Haste doesn't mean kicking people out of their homes - it does mean allocating the region for change. - and insisting that any changes - result in higher density outcomes.

Those neighbourhoods are nice precisly *because* they have no high desnsity housing.

This is perception.
When I was in Egypt I stayed in a resort - it was very high density - but also a very nice neighborhood. We had gardens, pools, beaches - well you get the idea.

If cities would reward high density housing with resort grade amenities - they could reclaim their air.

Sure - only one row of houses can really have a nice view - but the theory of shared resources says its better that MORE PEOPLE have this view than fewer.

And housing does not HAVE TO completely ruin the natural beauty - there are aztec designs - and a variety of old world looks which have natural appeal.

The most important task is to alter public opinion - that living near railroad tracks in tenement buildings is sexy.

Making sexy trains, and sexy tenement grounds is the solution to clean air.

AIK

Re:We need to fix the ones we have somehow.

[AmericanInKiev]

I don't know about a totally car free zone, trucks still need to be able to make deliveries for instance.

In Europe - Northern Germany (Oldenburg) to be exact - there are walking areas, and these rather samll delivery trucks come early in the morning.

I'm not sure if there's a sticker - a time - or a general understanding - but I didn't see unnecessary vehicles - and it is clear that thetruck are on pedestrian turf - not the ither way around - which drastically changes the way they are driven

AIK

Re:Theory and Implementation

[Marxist Hacker 42]

I agree, the liars can cause a problem, but this is an implementation issue, not a system issue.

That's kind of like saying that Microsoft allowing executable code in e-mail messages is an implementation issue, not a system issue. Idealy, the system should be engineered from the start to prevent parasites and viruses from arising to begin with. If it doesn't, then the system needs to be scraped and replaced with something better, just as many of us scrap Windows in favor of Linux.

If an entity can not create more value then it consumes it should die. This goes for both companies, ideas and people. Harsh, but that is the nature of the system. It is also the way of natural selection.

I would point out that if this is the case, then civilization itself is way too much of a bother- the wrong people are dying off thanks to parasites and viruses in the system. If we're going to have natural selection anyway, why bother having a civilization that protects the weak (the parasites and viruses) while punishing the strong (those who actually do the work)?

Christopher Alexander - "The Timeless Way..."

[garyebickford]

Object Oriented programmers are already aware of the congruence. Architect Christopher Alexander's books "The Timeless Way of Building" and "A Pattern Language" were the inspiration for the patterns movement in OO programming.

There are important lessons in these books for both urban design and system design. Many architects and urban designers don't like Alexander because his approach is counter to the "power over" (top down) approach to urban design, but encourages supporting bottom up, "power with", design-while-building that is characteristic of vernacular architecture. The problem, as they see it, is that they can't start building until the design is complete - the support systems (as well as the permit process) require it. They're wrong in principle but right in practice, because that's how the support systems already in place work.

Even when designers try to emulate the style of a village, it is still not quite like the real thing that grew over decades or centuries. However, perhaps automation could empower a kind of collaboration and serendipity so that a naive group of 'users' could essentially grow a design.