Index Tracking : Genetic Algorithms for Investment Portfolio Selection

摘要: This project was concerned with passive portfolio selection using genetic algorithms and quadratic programming techniques. Searching a large universal set of shares for subset that performs well is intractable, so stochastic search method must be used. The algorithm generates the subsets, used to find both their performance proportion available capital should invested in each member company. Separate subpopulations are maintained on different processors Meiko Computing Surface, occasional migration genomes. strategy allows several differing threads pursued within separate subpopulations, encourages convergence global optimum, instead local optima.