Fareed Sheriff’s algorithm to optimize traffic flow could earn him a $250,000 prize. (Photo courtesy Fareed Sheriff)
Fareed Sheriff, a senior at Mills E. Godwin High School, got tired of hitting red lights on his way to school. He asked himself, “Why does that keep happening, and how can I fix that?” He decided to solve the problem, writing an algorithm that optimizes traffic flow by predicting the future.
Sheriff’s algorithm has earned him one of 40 finalist spots in the Society for Science’s 2021 Regeneron Science Talent Search. Considered the premier high school science competition, with almost 1,800 entries, it awards at least $25,000 to each finalist, with the top prize reaching $250,000. As a finalist, Sheriff will attend a virtual finals week this month where judges will decide the allocation of top prizes.
“We are looking for innovation and … groundbreaking thinking in terms of science, technology, engineering and mathematics,” says Society for Science CEO Maya Ajmera. Sheriff’s project stood out because it addressed a common problem and “has a lot of public health benefits,” such as reducing accidents and pollution, Ajmera says.
Sheriff explains that modern, real-time traffic systems “optimize in the present, with no consideration of the past or future.” Instead, his algorithm uses machine learning to analyze past traffic patterns and predict future traffic flow.
Sheriff uses chess as an analogy. He compares real-time systems to novice chess players looking for the biggest advantage in the next move, regardless of the past or future. His algorithm is like an experienced chess player, in that it “thinks multiple moves ahead and yields better results than real-time optimization,” he says.
Sheriff plans to study mathematics and computer science in college and wants to test other applications of his algorithm to solve more everyday problems. “Abstract concepts ... have surprising applications in the real world if you use them correctly,” he says.