A hybrid of adaptive genetic algorithm and pattern search for stock index optimized replicate

How to allocate the weights of stocks is an interesting technology in stock index optimized replicate. This paper proposed a hybrid algorithm of adaptive genetic algorithm and pattern search (AGA-PS) to find the optimal portfolio weights. In AGA-PS, the crossover probability and mutation probability are adjusted adaptively. The weight from adaptive genetic algorithm is as the search start point of pattern search. The experiment result of AGA-PS, which has smaller tracking error compared with GA and AGA, has shown that AGA-PS model is feasible and effective to the stock portfolio.