Evolutionary Car Controller using (NEAT) by Saber Fazliahmadi

نرم افزار مطب آنلاین
نرم افزار مطب آنلاین
14 بار بازدید - 7 ماه پیش - """ ***Author and Developer: Saber
""" ***Author and Developer: Saber Fazliahmadi*** Evolutionary Car Controller using NeuroEvolution of Augmenting Topologies (NEAT) ## Introduction This program evolves a neural network to control a car agent to navigate a track. ## NEAT Overview NEAT (NeuroEvolution of Augmenting Topologies) is an evolutionary algorithm that creates neural networks and evolves them through generations to solve tasks. NEAT starts with simple neural networks and complexifies them over generations.(Over time by adding new nodes and connections through Mutation or Coress-Over) # The networks are evaluated on a fitness function based on how well they perform the desired task. The top performers reproduce to populate the next generation. by adding NODES and CONNECTIONS through MUTATIONS to find an optimal topology. The car agents are the entities that NEAT evolves. Each car is equipped with sensors and actuators, and the neural network controls the car by processing sensor data and generating actuator commands. This Python program implements an evolutionary car controller using the NeuroEvolution of Augmenting Topologies (NEAT). NEAT is an evolutionary algorithm designed to evolve neural networks over generations. The purpose of this program is to evolve a Neural Network to control a car agent, guiding it to navigate a track effectively through the concept of "without crashing". The car agents are the system that NEAT is evolving. Each car has sensors and actuators that the neural network controls. The neural network takes the sensor data as input and outputs actuator commands. The population of Neural Networks has evolved over generations. 1. Networks are initialized and evaluated on the car control task. 2. The Top-Performing networks reproduce to populate the next generation. 3. Networks Mutate to add structural variations. 4. Repeat evolution until an optimal controller is found. ***The neural networks are evaluated on how well they navigate the track wit
7 ماه پیش در تاریخ 1402/11/04 منتشر شده است.
14 بـار بازدید شده
... بیشتر