City Skylines Evolve Just Like A Living Thing, Reveals Genetic Algorithm That Predicts The Rise Of Skyscrapers

Bustling urban areas grow following patterns that are common to all living things, uncovered a new study by Spanish researchers from the University of A Coruna.

The paper, published in March in the Journal of Urban Planning and Development, reveals that the increase of skyscrapers in a busy city mimics the pattern "of certain self-organized biological systems," ScienceDaily reports.

To get to this conclusion, the team looked to nature for inspiration and developed a genetic algorithm that simulates the vertical growth of a city skyline.

By using historical and economic data on any given urban area, their genetic algorithm estimates how many skyscrapers will be built there on a certain timeframe.

In addition to predicting the number of new tall buildings that will likely appear in a specific neighborhood, the algorithm can also pinpoint the areas where the skyscrapers will be located.

Architect Ivan Pazos, the lead author of the new study, explains the science behind the algorithm.

"We operate within evolutionary computation, a branch of artificial intelligence and machine learning that uses the basic rules of genetics and Darwin's natural selection logic to make predictions."
The newly-developed genetic algorithm is based on standard versions of genetic algorithms created for previous and unrelated studies and can "learn the growth patterns of urban districts" by tapping into the city's construction history, notes ScienceDaily.

The Spanish researchers have already tested out their algorithm on the Japanese the Minato Ward in Tokyo, a fast-growing metropolis that houses the headquarters of many prominent multinational companies, such as Mitsubishi, Honda, NEC, Toshiba, and Sony.

The Minato Ward was chosen "because it has been one of the fastest growing skylines over the last 20 years," the authors write in their study. However, the team points out that their genetic algorithm can be replicated on any given city to predict how many skyscrapers will arise there in the future.

The Minato Ward test, carried out in 2015, rendered a series of maps and 3D models of the Japanese city, which the team can now compare with the evolving skyline.

According to the researchers, their genetic algorithm has been spot-on in describing how the city would look in a year's and two years' time.

"The predictions of the algorithm have been very accurate with respect to the actual evolution of the Minato skyline in 2016 and 2017," revealed Pazos.

"Now we are evaluating their accuracy for 2018 and 2019 and it seems, according to the observations, that they will be 80 percent correct," Pazos pointed out.

As he explains, the algorithm can be used to anticipate how cities evolve, by feeding it a large number of possible scenarios, from which its selection system repeatedly chooses the best and most accurate options.