Genetic Algorithms: Playing God: Elitism vs Egalitarianism
Genetic Algorithms (GA) are a class of Artificial Intelligence that encode your model variables as chromosomes. And then you breed them over hundreds of generations mimicking evolution, using crossover and mutation of genes to improve the species. GA applications include optimizing airline schedules, discovery of new drugs and new metal alloys, optimizing capacity of transportation systems etc. But I don’t have the data to practice on these topics, so I use GA to optimize portfolio strategies for the 30 constituents of the Straits Times Index (STI) and compare their performance. The two GA approaches are (1) Elitism where parents compete in a tournament (like the Hunger Games) and only high-performance parents are allowed to breed and in each generation the children with good performance become parents, and the weak performance offspring are killed. (2) Egalitarianism where any compatible parents are allowed to marry and produce offspring. The performance of strategies 1 and...