27 Jan 2005
CS 3243 - Heuristics
39
Genetic algorithms
nA successor state is generated by combining two parent states
n
nStart with k randomly generated states (population)
n
nA state is represented as a string over a finite alphabet (often a string of 0s and 1s)
n
nEvaluation function (fitness function). Higher values for better states.
n
nProduce the next generation of states by selection, crossover, and mutation