Cumulative Prospect Theory oriented algorithms for portfolio selection: an empirical application using swarm intelligence

Abstract

Since the late 1970s, academia has pioneered a radical new approach to the study of the strategies and decision-making processes of economic agents, introducing behavioral variables into the models before eventually rejecting the rational investor paradigm. Tversky and Kahneman were the founders of behavioral finance and are credited with creating Prospect Theory. In this study, the Gray Wolf Optimization algorithm and Particle Swarm Optimization algorithm, two evolutionary methods for portfolio selection, are used in conjunction with the Cumulative Prospect Theory model. A portfolio optimization application will be carried out using daily equities data from the FTSEMIB. Investigations have been conducted in order to assess which algorithm performs better. PSO algorithm has proven to perform a better selection according to the benchmarks, three financial indicators and the magnitude of violations of the problem constraints. Ultimately, further investigation of the CPT model parameters is done.