Comparison of Genetic Algorithms with Conjugate Gradient MethodsGenetic algorithms for mathematical function optimization are modeled on search strategies employed in natural adaptation. Comparisons of genetic algorithms with conjugate gradient methods, which were made on an IBM 1800 digital computer, show that genetic algorithms display superior performance over gradient methods for functions which are poorly behaved mathematically, for multimodal functions, and for functions obscured by additive random noise. Genetic methods offer performance comparable to gradient methods for many of the standard functions. |
Other editions - View all
Common terms and phrases
adaptive control associated inversion pattern average behavior best string better string chosen randomly Conjugate Gradient Algorithms conjugate gradient methods control systems convergence rates corresponding alleles crossover and inversion direct search algorithms effectiveness fixed step Fletcher-Reeves method four strings four subpopulations function evaluations required function evaluations taken function optimization problems Function Value Attained function value level genetic algorithms genetic methods gradient mutation highest function value history vector Index Squared indicated integer kind of cross-over last adaptation m₁ strings methods of mutation mutation method mutation routine Newton Raphson number of function number of strings Optimal Control optimum pivot points probability vector Rastrigin rate of convergence Repeated Peaks replaced by string replaced the worst required by Version reset interval q Schumer's method search strategies employed second level adaptation Spherical Contours standard deviation strings were chosen Table 2a Test Function uniform random University of Michigan utility vector version II system Version IV's