Multi-Objective Optimization of Cancer Chemotherapy Using Swarm Intelligence

Portfolio optimization is a problem that lends itself naturally to multiobjective approaches, e.g., aimed to maximize the return of the investment, simultaneously minimizing the risk. The selection of an actual portfolio requires exercising a decision-making process on the set of efficient solutions thus obtained. In this work we consider the case in which knowledge of this selection criterion is available, and used within the optimizer. We use Sharpe’s index, a measure of excess return per unit of risk, for this purpose. It is shown that a multi-start single-objective evolutionary algorithm based on this index can provide a better coverage of the relevant regions of the Pareto front than state-of-the-art multiobjective evolutionary algorithms. An extensive experimental analysis is conducted using real data from a Latin American stock exchange.