MSc Thesis Defence by Paul Seward
MSc in Mathematical and Computational Sciences Thesis Defence
Presenter: Paul Seward
Title: 鈥淎dvancing Particle Swarm Optimization through Deep Reinforcement Learning鈥�
Metaheuristic algorithms like Particle Swarm Optimization (PSO) are widely used for complex optimization but often suffer from premature convergence and stagnation. This thesis integrates Deep Reinforcement Learning (DRL) with PSO to develop intelligent, adaptive strategies that address these inefficiencies. We model the swarm of particles as an agent-controlled environment and explore different DRL architectures that enable discrete, continuous, and subswarm-based interventions. Extensive experiments on the IEEE CEC鈥�13 benchmark suite demonstrate that hybrid DRL-driven policies consistently outperform standard PSO, highlighting the value of multi-action and parallel interventions. This research provides new insights into adaptive metaheuristics and outlines promising directions for intelligent optimization.
Date and Time: May 14 at 9:00 am via Web Conference
To receive the link to the public presentation, please email the graduate studies coordinator at gsc@upei.ca
Everyone is welcome to attend.
