Premium
Genetic algorithms: An evolution from Monte Carlo Methods for strongly non‐linear geophysical optimization problems
Author(s) -
Gallagher Kerry,
Sambridge Malcolm,
Drijkoningen Guy
Publication year - 1991
Publication title -
geophysical research letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.007
H-Index - 273
eISSN - 1944-8007
pISSN - 0094-8276
DOI - 10.1029/91gl02368
Subject(s) - monte carlo method , quasi monte carlo method , computer science , linearization , genetic algorithm , algorithm , monte carlo method in statistical physics , monte carlo molecular modeling , hybrid monte carlo , monte carlo integration , mathematical optimization , nonlinear system , mathematics , markov chain monte carlo , machine learning , physics , statistics , quantum mechanics
In providing a method for solving non‐linear optimization problems Monte Carlo techniques avoid the need for linearization but, in practice, are often prohibitive because of the large number of models that must be considered. A new class of methods known as Genetic Algorithms have recently been devised in the field of Artificial Intelligence. We outline the basic concept of genetic algorithms and discuss three examples. We show that, in locating an optimal model, the new technique is far superior in performance to Monte Carlo techniques in all cases considered. However, Monte Carlo integration is still regarded as an effective method for the subsequent model appraisal.