Algorithme génétique

Un article de Wikipédia, l'encyclopédie libre. Origines[modifier | modifier le code] La popularisation des algorithmes génétiques sera l'œuvre de David Goldberg à travers son livre Genetic Algorithms in Search, Optimization, and Machine Learning[1] (1989). Ce livre est encore édité aujourd'hui. Présentation[modifier | modifier le code] Analogie avec la biologie[modifier | modifier le code] Terminologie commune aux deux disciplines[modifier | modifier le code] Les algorithmes génétiques étant basés sur des phénomènes biologiques, il convient de rappeler au préalable quelques termes de génétique. Les organismes vivants sont constitués de cellules, dont les noyaux comportent des chromosomes qui sont des chaînes d'ADN. On utilise aussi, dans les algorithmes génétiques, une analogie avec la théorie de l'évolution qui propose qu'au fil du temps, les gènes conservés au sein d'une population donnée sont ceux qui sont le plus adaptés aux besoins de l'espèce vis-à-vis de son environnement. Sélection

Genetic algorithm The 2006 NASA ST5 spacecraft antenna. This complicated shape was found by an evolutionary computer design program to create the best radiation pattern. Genetic algorithms find application in bioinformatics, phylogenetics, computational science, engineering, economics, chemistry, manufacturing, mathematics, physics, pharmacometrics and other fields. Methodology[edit] In a genetic algorithm, a population of candidate solutions (called individuals, creatures, or phenotypes) to an optimization problem is evolved toward better solutions. Each candidate solution has a set of properties (its chromosomes or genotype) which can be mutated and altered; traditionally, solutions are represented in binary as strings of 0s and 1s, but other encodings are also possible. A typical genetic algorithm requires: a genetic representation of the solution domain,a fitness function to evaluate the solution domain. Initialization of genetic algorithm[edit] Selection[edit] Genetic operators[edit] Termination[edit]

Utilizing Genetic Algorithms to Identify Potential Software Performance Opportunities – Blogs Tech Project #1: Utilizing Genetic Algorithms to Identify Potential Software Performance Opportunities In these blogs, I would like to discuss some of our failed technical projects at Intel in order to share some of the lessons we have learned. I am a believer that you learn just as much from your failures as from successes…so in these projects we learned a lot. This tech project ended back in 2004 and was absolutely fascinating. The parent and newly created children binaries were then compared in a bout of performance. One example of the random changes which was applied was a temporal locality hint which can be applied to loads or stores on the Itanium architecture. a) ld8 [r23] // Before temporal locality hint is applied as a mutation b) ld8.nta [r23] // After a temporal locality hint is applied as a mutation If this change increases the performance of the binary then that binary will win in the next contest and become the new parent binary…moved to the next round.