Dynamic hill climbing

Michael de la Maza and Deniz Yuret (1994) ( PS )
Dynamic hill climbing. AI Expert.

Abstract:

In this article we describe a new optimization algorithm, called dynamic hill climbing (DHC), that draws on ideas from genetic algorithms, hill climbing, and conjugate gradient methods. On DeJong's five function test suite, a well known set of functions that has been widely used to study optimization algorithms, dynamic hill climbing outperforms genetic algorithms and simulated annealing, often by a wide margin. DHC is currently being used in a medical imaging system that helps brain surgeons to plan and execute operations.