The program running.
The great algorithm of natural evolution and the problems that traffic in all cities represent (economic losses, environmental pollution, and social wellbeing, among others).
What it does
It generates a random distribution of cars which serve to test traffic lights from an intersection that are controlled by a set of instructions, and these instructions are constantly optimized with a genetic algorithm.
How we built it
We developed a program with processing that uses a set of algorithms made by us and inspired by previous works from other researchers.
Challenges we ran into
We had to implement an AI system for the cars, which took us more than we expected. And in the final step, when we went from simulating one by one to simulating a whole population.
Accomplishments that we're proud of
The level of fitness that the individuals achieved after being exposed to constant evolution for many generations.
What we learned
How to implement a genetic algorithm to optimize a function. Also, we learned how to work with the language processing. How to make a simple AI, physics engine and motion simulation for animated cars.
What's next for Smart Traffic
Implementing it in a system of multiple intersections with more lanes. Applying it to real world traffic lights. Instead of just evolving an instruction, making the algorithm to evolve the network.