Global optimization, search and machine learning methods based on genetic algorithms

摘要

Global optimization, retrieval and machine learning method based on genetic algorithmGlobal optimization based on genetic algorithm of the present inventionSearch and machine learningStep 1 based on the objective function f (x) of the structure of matterA problem substance was edited as a chromosome of genetic algorithm.Process 2 is initialized with automatic calculation or input, andBased on the evaluation for the objective function f (x)The genetic operator of the acquired traits of the present invention and the "use and abolition operator" of the present invention provide a new generation (k generation) group GK = {p1KP2k,...PS3 is optimized for iteration of PS3.The step 4 of the optimum solution of the subject is output. The present invention combines the natural law of "genetics of acquired traits" and "disposable operators" in the evolution of Lamarck's evolution within the natural law of "epigenetics" and Darwin's evolutionary theory.Simplify the genetic algorithm structure. Overcoming many technical flaws in conventional algorithmsIncrease the efficiency.Improving the global optimization and sustainability of late evolutionMore global optimizationA more preferred solution is obtained for the problem of search and machine learning.The present invention combines the natural law of "genetics of acquired traits" and "disposable operators (evolutionary and dissemination operators)" in Lamarck's evolutionary theory within the natural law of modern "epigenetics" and Darwin's evolutionary theory, "genetic survival" Simplify the luggage structure.Overcome many technical deficiencies in traditional algorithms, increase their efficiency, improve global optimization and sustainability of later evolution, and more favorable solutions for more global optimization, search and machine learning problemsDiagram