CSCE 315 Lecture 18

From Notes
Jump to navigation Jump to search

« previous | Friday, March 1, 2012 | next »


Optimizing Search Functions

Iterative Improvement

Guess at first, then adjust for more accurate results

Hill Climbing

Start at a random point. Choose the best of 3 random points. Repeat until no improvements

Gradient Ascent

Take gradient of function, then ascend a step in that direction.

Simulated Annealing

"heat up" the entropy of random points. As it cools down, settle on maximum points. Temperature allows accepting of worse states at first

Genetic Algorithms

State = "chromosome" and optimization function = "fitness". Create "generations" of solutions, "kill off" the worst solutions, and "cross-breed" the best to get better state