Name two ways animals adapt?
What is the Fitness Function formula for evolutionary robotics
Explain Evolutionary Robotics
Evolutionary robotics uses evolutionary methods, especially genetic algorithms, to automatically design robot controllers for engineering and scientific purposes.
How is an evolved genotype evaluated in evolutionary robotics
By decoding the robot’s network, body, sensor, and motor parameters, building the robot, testing it in an environment, scoring its fitness on a task, and then selecting, reproducing, and mutating the best designs.
What is the difference between simulation and physical robots in evolutionary robotics?
Simulations are fast and repeatable, while physical robots are realistic but costly and time-consuming.
What is evo-devo in evolutionary robotics?
An approach that evolves both a robot’s body and brain together.
What is swarm robotics in evolutionary robotics?
An approach where multiple robots cooperate as a group using evolved controllers.
What is soft robotics in evolutionary robotics?
An approach that applies evolutionary methods to flexible, soft-bodied robots.
Name the 3 types of machine learning
Explain Reinforcement Learning
Interactions with an environment and taking actions through experience often relating to reward factor.
What does Q(s, a) estimate in reinforcement learning?
The discounted cumulative reward for taking action a in state s and then following the current policy.
What is the optimal action in a state according to the Q-function?
The action with the highest Q-value, written as argmaxₐ Q(s, a).
Why are neural networks used in reinforcement learning?
Because the Q-table can be too large or infinite.
What does a Q-network learn?
It learns the Q-function by mapping states to action Q-values.
What is one benefit of using a Q-network?
It can generalise across similar states.
