top of page

Modeling a City's Freeway Network

Central Business Districts, Functionally Unified

What does it mean for a business district to be unified, continuous, not split-in-two? Midtown and Downtown Atlanta are not considered unified by any standard. Neither in New York, with its distinctions between Downtown, Midtown, and even Midtown South. Many cities have more than one business district at their centers, but for my purposes, a business district is functionally continuous if the block-by-block proportion of commuter origins is consistent.

By observing existing business districts, real estate prices, and talent agglomerations, it may be undesirable to design business districts with more than a 5 minute difference from its first freeway exit to its last freeway exit. For subway-oriented downtowns, this appears to translate in the real world. A 5 minute ride from Midtown Manhattan's northern-most subway stations puts one in Midtown South. Because subway trains need to stop frequently, 5 minutes in downtown moves commuters less than 2 miles. Business districts which try to spread themselves further than five minutes quickly split apart or fray at the edges. Defining freeway speeds could be important for defining how large downtowns should functionally be.

Atlanta's downtown connector freeway moves traffic in rush hour not much more than 20 miles per hour, resulting in two functionally separate business districts, Midtown and Downtown, each 1.5 miles long. At 90 mile-per-hour, Atlanta's downtown connector freeway can produce a unified central business district that is 7.5 mile long. A consistently 60 mile-per-hour freeway, however, could only produce a business district 5 miles long. Currently, central city freeways of major cities are congested and produce slower speeds than many subway systems, especially in rush hour, and produce downtowns which are physically smaller.

How Fast Can autonomous Freeways Go?

There may be practical limits with freeway speeds as they begin to supersede 100 miles-per-hour with variables like the wind and even mild bumps in the road surface beginning to cause hard-to-manage jumps in jerk and acceleration above one meter-per-second-squared (the industry standard used for mass transit), making rides uncomfortable. Faster freeway speeds may also require excessive energy consumption compared to air planes, not to mention longer reaction times that could conversely require a decrease in traffic flow compared to a slower freeway with quicker reaction times. These may be some of the reasons that futurists dating back to the 1930s dreamed of freeways with top speeds of120 miles-per-hour. If cities prepare for the possibility of 120 mile-per-hour freeways and plan out a metropolitan region by spacing out future travel times, 120 mile-per-hour freeways may still be the best assumption we have.

Limiting Factors

A downtown freeway can only have so many exits before becoming functionally crowded for merging. Exit routes can only travel so far from the main freeway before they become time-inconvenient. Exit routes must be crossable for pedestrians and local traffic, lowering each exit's flow potential.


These are powerful limiting factors. My metro area-wide prioritization algorithm systemically targets exit routes to not exceed half of their capacity during peak traffic could give cities a kills-multiple-birds solution to handle the moment-by-moment variance of arriving traffic for each freeway exit. In addition to this, variable pricing of traffic would be a powerful policy tool which would relieve connected traffic systems from leaning too heavily on selective congestion alone. For reasons described in my Mass Transit Model, exit routes with more than four lanes require exhausting merging patterns which themselves take up exponentially more room. All of these factors create limits for what a perfect downtown traffic structure can look like.

How Atlanta compares with Competitors

Without adjusting for the angle of its embankments, Atlanta's downtown connector freeway (I-75/85) measures as capable of going no faster than 60 miles-per-hour in some parts, without pushing curve acceleration past one meter-per-second-squared.


For a more competitive design, Atlanta may need to re-align three sections of its freeway (the north and south ends of the "Grady Curve" and the northern formation of the downtown connector). If Atlanta secured these rights-of-way in advance, the realignment of those sections would allow Atlanta's downtown connector freeway to process traffic at a modest 90 miles-per-hour, nearly doubling the speed and the potential parameters of its future central business district. This, combined with a 90 mile-per-hour southern half, would result in a central business district which effectively spans the entire downtown connector freeway---approximately 5.45 minutes would separate the freeway exits proposed in my Mass Transit Model from Midtown to the northern extreme and East Point to the southern extreme, and 4.69 minutes would separate their respective freeway entrances, making the furthest routes functionally 5.055 minutes apart.

Assuming the interchange between Atlanta's downtown connector freeway and I-20 is redesigned, a widening of certain parts of Atlanta's downtown connector freeway south of Midtown would make I-20 an co-equal commuting branch to I-85 and I-75, and increase arriving traffic volume by 50% while increasing the number of neighborhoods conveniently close to Atlanta.

Currently, human drivers are unable to coordinate merging effectively, so it's pointless to build freeways more than 6 lanes wide in each direction. Extra lanes create more traffic from all the necessary merging. Computer-guided traffic, by contrast, could enable near-perfect merging and allow freeways to carry more freeway lanes (use wisely), so long as there is enough right-of-way committed for merging at strategic points. Merging patterns could be important for the rights-of-way cities need to secure for automated traffic. All of Autonomous Symmetry's designs are predicated on these merging equations.

choice for cities: self-surgery or die


The busiest central business district in the world is Midtown Manhattan. Approximately 800,000 people reportedly work in Midtown Manhattan. This does not include tourists, students, and non-working residents. Meanwhile, approximately 150,000 people work in Downtown Atlanta. With a traffic system which delivered 800,000 vehicles every day into the central business district from the freeway alone, such an expanded downtown would likely have more than a million people in commuting in and out each day. Considering the possibilities of mini-buses and autonomous mass transit, it is not unrealistic to think about the expanded downtowns of Atlanta, Houston, Los Angeles, and Dallas as eventually having millions of daily commuters by the end of the 21st century.

Cities like Dallas, Los Angeles, and Singapore share Atlanta's advantageous situation for a powerful central city freeways, and strong surrounding networks.

While other cities could theoretically engineer ambitious systems for their central business districts, three existing central business districts stand out to Autonomous Symmetry---Atlanta (I-75/85), Los Angeles (I-110), and Dallas (I-30/I-35/TX-183)---for their ability to potentially direct so much traffic into their expanded centers without need for deeply controversial levels of imminent domain and financial cost. Beijing (2nd Loop), Denver (I-25), and Austin (I-35) also have ideal amounts of right-of-way around their central city freeways and with more ambitious changes, they too would have top-tier central city traffic structures for all time. Other cities, like Houston, Boston, and Guangzhou, could​ achieve timeless designs with clever-enough changes if they are ambitious enough.

Equally important, these cities have the ability to give their core districts a pedestrian experience unmatched by any of today's major cities. At its core, my Mass Transit Model is a mathematical proof-of-concept meant to guide and inform a city seeking to avoid planning mistakes before automated traffic becomes a reality.

Meanwhile, cities with responsive zoning, low costs of living, and flexible rights-of-way for automated traffic could have crucial advantages as industries die, migrate, and blossom. Like Dallas and Los Angeles, Atlanta has a great deal of residential growth occurring around its central freeway, meaning such decisions could likely become more politically difficult the longer such planning decisions are held off. The wrong things keep being built in the wrong places---valuable projects adjacent to the wrong infrastructure and in crucial rights-of-way. Residential zoning of largely vacant neighborhoods on the west and south sides of Atlanta could, if not changed, reduce the commercial and mixed-use land available to Atlanta's future central business district, possibly preventing an aesthetically unified district from forming. Potential pedestrian grids, if not planned in advance, will be bogged down with their responsibilities connecting parking garages and other structures which could have been foregone or better-placed.

Automated traffic could ensure freeways can go faster, consistently and reliably, but only only as fast as designed. Some freeways are straight enough to avoid curves which cause acceleration for passengers, but others could need to go slower. Slower freeways could connect fewer neighborhoods to be conveniently close enough, resulting in a less competitive central business district.

Freeway Speeds

If Atlanta's automated freeways were only able to carry traffic comfortably at 60 miles per hour, Atlanta could have 500 square miles of land within a 30-minute ride to its central business district (including the extra time to arrive off the freeway or connect with autonomous mass-transit). With 90 mile-per-hour speeds, this increases up to 1250 square miles----more than twice the land within 30 minutes of the central business district.

The hilly landscape of North Georgia has caused some of its freeways to be only as straight as they've needed to be. By contrast, the flat geography of cities in central Texas has made it easier to build neatly straight freeways. Dallas’s freeways are largely well-designed for 100 mile-per-hour traffic. Most of Atlanta's freeways could also carry 100 mile-per-hour traffic, but there are kinks which force traffic to slow down, sometimes as slow as 60 miles per hour.

If Atlanta and Dallas were eventually populated with 5,500 people per square mile, Atlanta's 60 mile-per-hour freeways would put 2,500,000 people within 30 minutes of downtown, while 90 mile-per-hour freeways would put 7,500,000 people within 30 minutes of Dallas's downtown.

If Atlanta's freeways are not realigned in certain parts, it may be difficult for Atlanta to compete with cities like Dallas.

Single-Center Cities vs Hub-and-Spoke

In growing cities, traffic could continue to max out freeway flow capacity, even if cars are ubiquitously connected and controlled and freeway traffic moves at top speeds. But with connected and controlled traffic, cities could be tasked to choose which pain points (selective congestion, variable pricing) to use in their metro-area prioritization algorithms. These pain points could intertwine with market forces and autonomous mass transit to determine how cities reshape themselves and grow.

Cities with singular centers today seem to grow slowly under natural limits. Cities like New York struggle to stretch and create ever-denser-development around impressive, overburdened mass transit and freeway infrastructures, a sign of New York City's magnetism and limits. Unlike New York, metropolitan areas with multiple hubs and spokes have no trouble growing outward, but it is harder to create density or valuable talent agglomerations this way. Most modern cities have transformed themselves towards hub-and-spoke systems, but traffic chokes much of their potential. Dallas and Atlanta are two large metropolises which have achieved strong hub-and-spoke models, yet relatively small downtowns. With automated traffic and autonomous minibuses, the ambitious freeway structures of Dallas and Atlanta could conversely allow them to build two of the largest central business districts on Earth if desired.

Regional Business Districts

Currently, business hubs tend to spread themselves out by roughly 10 to 30-minute increments. In rush-hour Atlanta, it takes an average of 30 minutes to travel from Dunwoody to Cumberland, 17 minutes from Cumberland to Midtown, and 15 minutes from Midtown to Buckhead. In Dallas, it takes 23 minutes from Downtown to Addison, 35 minutes from Downtown to Fort Worth, but only 10 minutes from Addison to Legacy.


There’s actually an underlying pattern here, but it requires measuring these distances at night: business districts in Dallas and Atlanta, once stripped of their traffic congestion, are spaced out in segments which mostly take 10 minutes to get from one business center to the next. I see this trend in other cities as well, including cities which largely depend on subway systems rather than freeways. This suggests that business districts naturally establish themselves in 10-minute increments and are stretched apart by traffic congestion overtime.


10-minute increments


While 10-minute increments for regional business districts seems like a good place to start, there are caveats. Where central business districts are close, adjacent business hubs may risk being too close to the center’s newly strengthened gravity, especially given the ability of connected traffic systems to prioritize traffic in and out of a central city. Places like Cumberland and Buckhead in Atlanta, or Addison in Dallas, may find themselves shifting to greater roles in high-value retail and dense residential land uses while holding onto proportionally smaller roles as office hubs.

90 mile-an-hour Freeways could be the foundation of mass transit

With 90 mile-per-hour freeways, business hubs like Legacy and Fort Worth in Metro-Dallas, or Northern Alpharetta in Metro-Atlanta would be approximately 20 minutes outside of their respective central business districts. These places could have more distance to resist the economic gravities of their central business districts and flourish in their own regional supremacy.

bottom of page