Lead PI: Professor Michael O’Neill
This work package focuses on strategic asset allocation in both equity and fixed income asset classes. Challenges in this area include huge datasets and the dynamic nature of financial asset prices. Model induction technologies such as GP offer particular utility for this problem as they can embed human-knowledge / mandate constraints and can output human-readable asset selection rules / models.
GP falls under the banner of evolutionary algorithms (EAs), which draw inspiration from the processes of biological evolution to breed solutions to problems. An EA typically commences by creating an initial population of potential solutions and these are iteratively improved over many ‘generations’. In successive iterations of the algorithm, fitness-based selection takes place within the population of solutions. Better solutions are preferentially selected for survival into the next generation of solutions, with diversity being introduced in the selected solutions in an attempt to uncover even better solutions over multiple generations.
Professor O’Neill and Professor Brabazon have recently authored a major book on the application of grammatical evolution (GE) and other Natural Computing algorithms to the domain of Financial Modelling (Brabazon and O’Neill, 2006b). Building on this expertise, it is proposed to continue to extend the development of GE for application in dynamic financial environments.