SEE THE SOURCES USED FOR THE GLOSSARY DEFINITIONS HERE
Agents perceive and act upon the environment they are situated in. Each action results in a new state of the environment.
Evolutionary algorithms use evolution as an inspiration for solving computational problems. They share the following basic features:
- A population of individuals which are potential solutions to a given problem,
- A measure of fitness assigned to individuals that determines the quality of the solution,
- Selection of individuals based on their fitness to generate new individuals through the use of operators.
Examples include genetic algorithms, genetic programming, and evolution strategies.
STRUCTURE OF AN EVOLUTIONARY ALGORITHM
g ← 0 initialize P(g) evaluate P(g) WHILE NOT (termination condition) DO g ← g + 1 select P(g) from P(g-1) apply variation operators to P(g) evaluate P(g) END // P = population, g = generation
A node is expanded by applying each legal operator to the state the node represents, generating a new set of nodes in the process.