A Multi-Objective Evolutionary Algorithm for Portfolio Optimisation

The use of heuristic evolutionary algorithms to address the problem of portfolio optimisation has been well documented. In order to decide which assets to invest in and how much to invest, one needs to assess the potential risk and return of different portfolios. This problem is ideal for solving using a Multi-Objective Evolutionary Algorithm (MOEA) that maximises return and minimises risk. We are working on a new MOEA loosely based on Zitzler’s Strength Pareto Evolutionary Algorithm (SPEA2) [20] using Value at Risk (VaR) as the risk constraint. This algorithm currently uses a dynamic population in order to overcome the problem of loosing solutions. We are also investigating a dynamic diversity and density operator.