A memetic algorithm for cardinality-constrained portfolio optimization with transaction costs

Rubén Ruiz-Torrubiano, Alberto Suárez

Abstract: Highlights•Novel memetic algorithm for optimal portfolio selection.•Proposed framework considers transaction costs, cardinality and other real-world constraints.•The combinatorial and the continuous optimitzation aspects of the problem are handled separately.•Adaptation of the RAR crossover operator for an extended set encoding.•We find that using certain regularization mechanisms results in more efficient portfolios.•Ignoring transaction costs results in inefficient portfolios out-of-sample.