The efficient management of assets to cover a firm’s liabilities over a multi-period horizon is a relevant challenge for many financial companies. Even in its deterministic version, this problem is complex since managers have to make difficult decisions about their asset portfolio each period. With the goal of maximising the expected terminal wealth in a scenario under uncertainty, this paper proposes a novel simheuristic approach that integrates Monte Carlo simulation at different stages of a Genetic Algorithm. Our approach is capable of generating effective solutions to the considered problem in relatively short computational times. Moreover, our simheuristic is enriched with several “smoothing” techniques that enhance the attractiveness for managers of the generated solutions, so they can be effectively employed in real-life applications. A series of computational experiments, including the use of advanced evolutionary strategies, illustrate these concepts and justify the advantages of including simulation in financial optimisation problems under uncertainty.
- asset and liability management
- genetic algorithms