Congratulations to INET Oxford Doctoral Student Aymeric Vié, who won a Best Paper Award at the Genetic and Evolutionary Computation Conference (‘GECCO’ – a prestigious conference in evolutionary computation).
Aymeric, who is a member of INET Oxford’s Complexity Economics programme, was awarded the prize for his paper ‘Population network structure impacts genetic algorithm performance’.
His DPhil thesis focuses on evolutionary market ecologies: models of the financial markets where agents have various financial strategies and learn over time to improve their performance. This learning is modelled using evolutionary algorithms: a subset of machine learning that simulates natural evolution.
In financial markets, the network structure between economic agents is likely to change their interactions, opinion dynamics, returns and economic patterns; this paper explores the impact of network structure over the performance of these evolutionary algorithms. The results showed that this network structure had a significant impact on the algorithm outcomes, and also their performance. This finding will allow for a more informed modelling of financial learning, and opens new perspectives in evolutionary computation.