genetic algorithms: represent knowledge as groups of characteristics. are based on logic. develop solutions to particular problems using inheritance, crossover, and mutation. seek to emulate a human expert's way of solving problems. do not work for most problems.