Expected mean return—standard deviation efficient frontier approximation with low‐cardinality portfolios in the presence of the risk‐free asset

In the expected mean return, standard deviation portfolio selection problem, first step is usually to derive set of efficient portfolios, which in space objective function values represented by frontier. With modern methods and software, it an easy task even for thousands assets provided that problem continuous. However, investors often introduce requirement limit number portfolios (portfolio cardinality). The resulting mixed-integer quadratic formulations are computationally much more complex. this work, we assume besides risky assets, risk-free asset available investor, short selling not allowed. Since case frontier cannot be directly derived a solver, propose naive but intuitive heuristic approximate presence asset. contrast general-purpose evolutionary heuristics, exploit underlying mechanism composition. We show numerical experiments large-scale instances works well, compared state-of-the-art multiobjective optimization algorithm. Moreover, produces remarkably limited cardinalities.