资 源 简 介
This Project used Genetic Programming to arrive to a function close enough to a target function given by a training data.
The program takes in Training data and generates random population of solutions(generation consists of functions/individuals). The program then evaluates the fitness (closeness) of each function to the target function (using training data). Based on those fitness values a new generation (evolution) is created using Crossover and Mutation operations on selected individuals (functions). The evolution is repeated while keeping track of the function with best fitness value until we meet a desired fitness value or meet the maximum number of generations allowed.
The following are settings configurable before the run of the program:
- Population size
- Max depth of tree representing each function/individual
- Crossover operation frequency
- Mutation operation frequency
- Method of Generation (Full, Grow, or Half and Half)
- Method of Selection (Tournamen
