Skip to main content

Financial Evolutionary Computing

Clack's research in the area of financial evolutionary computing has for the most part explored financial applications of genetic algorithms (GAs) and genetic programming (GP). This work contributed to evolutionary computing technology as follows:

  • Gene dominance during crossover (providing a form of genetic memory for GAs).
  • Improved efficiency and expressivity of GP technology via strong typing constraints, polymorphism, recursion, and lambda abstractions.
  • Improving the robustness of GP solutions via the maintenance of phenotypic diversity in the population, with volatility-adjusted fitness and voting mechanisms. This work demonstrated a 300% performance improvement over standard mechanisms for investment portfolio optimisation, and has been shown to have superior performance to Support Vector Machines.
  • Detecting the problem of anomalous risk-switching in multi-objective genetic programming. In particular, the work on multi-objective genetic programming in finance identified an important instability problem with efficient frontiers (a critical issue for automated investment).

A 2007 GECCO paper (Evolving robust GP solutions for hedge fund stock selection in emerging markets, Yan and Clack, 2007) won a Best Paper prize and Clack subsequently received over 20 invitations to give talks and run workshops at national and international conferences. This included invited talks at GECCO for three successive years (2008-2010). He received invitations to collaborate with international research groups (e.g. at Strasbourg, Complutense, and Budapest) and over 30 invitations to advise financial institutions: his new techniques for dynamically-optimised trading and investment were thereby transferred widely into industry and influenced the uptake of evolutionary computing approaches to automated trading and investment (e.g. by investment funds such as SIAM Capital)