MULTIPERIOD PORTFOLIO OPTIMIZATION WITH MULTISTAGE STRATEGY UTILIZING PARTICLE SWARM OPTIMIZATION

• Main Author: Arkaputra Azis, Rheznandya
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84220
• Institution: Institut Teknologi Bandung

Abstract Portfolio optimization is a common challenge in modern finance, especially considering complex and dynamic market uncertainties. In this study, we develop a novel approach to optimize portfolios over multiple periods using a multistage strategy integrated with Particle Swarm Optimization (PSO) algorithm. This multi-stage strategy views portfolios over multiple periods as a long-term process with different investment phases that reallocate assets depending on expected market conditions. This study discusses two common constraints in investment practices: bounded control, which limits asset allocation, and no-bankruptcy, which ensures portfolio sustainability without risk of bankruptcy. These constraints increase the complexity of multi-period portfolio optimization models, which can be addressed using PSO. PSO was chosen due to its ability to handle nonlinear and multidimensional optimization problems, as well as its flexibility to find optimal solutions in complex search spaces. Research shows that PSO effectively optimizes multi-period portfolios with multi-stage strategies, allowing for the management of large asset portfolios, long time horizons, and dynamic investment return schemes. Furthermore, the study sheds light on how constraints can affect investors’ return expectations, providing valuable insights for financial practitioners and academics. Thus, the use of PSO in this context not only efficiently optimizes asset allocation, but also provides a solution that addresses the complexities of multi-period portfolio optimization while taking into account relevant constraints in daily financial practice.