This Unity project demonstrates the usage of a genetic algorithm to calculate the required forces to successfully shoot a basketball into the basketball hoop from any position in the basketball court.
Genetic algorithms involve generating a population of random solutions, and evolving these solutions according to a specific goal. These solutions are represented as chromosomes. In this scenario, chromosomes represent solutions of basketball shots. Chromosomes include two parameters, namely an upward force and a forward force.
The fitness of chromosomes are evaluated by processing a basketball shot with its two force parameters after a specified duration. Fitness is a term used to describe the strength of the solution. The fitness of chromosomes are evaluated based on two factors, namely:
- The closest distance between the basketball and the basket during the evaluation; and
- If the basketball had successfully been scored in the basketball hoop.
The evolution of a chromosome population involves the following three phases:
- Selection - selecting chromosomes with mostly high fitness measurements (roulette method)
- Crossover - breeding new chromosomes that were selected during the selection phase
- Mutation - adding a small probability of changing parameters of chromosome offsprings that were bred during the crossover phase
By continuously evolving a population of chromosomes, the solutions slowly converge towards the specified goal. Therefore, the basketball shots slowly converge towards successfully being shot into the basketball hoop.
First Generation
100% Score After 17th Generation
Computational Intelligence: An Introduction by Andries P. Engelbrecht