Nur Iffah Binti Mohamed Azmi, <br/><br/>
Novel particle swarm optimisation approach integrating lévy flight and doppler effect for pid parameter optimisation in dc motor systems / 
Nur Iffah Binti Mohamed Azmi 
 - xviii, 192 pages :  illustrations ;illustrations ;  30 cm. +  1 CD-ROM 
<br/><br/> Faculty of Manufacturing & Mechatronics Engineering Technology 
<br/><br/>Thesis (Doctor of Philosophy) -- Universiti Malaysia Pahang – 2025 
<br/><br/>Includes bibliographical references 
<br/><br/> Optimisation plays a critical role in solving complex engineering problems, particularly within constrained optimisation and control system tuning domains. This research proposes an enhanced swarm intelligence algorithm (Particle Swarm Optimisation with Lévy Flight and Doppler Effect (PSO-LFDE)) designed to address the limitations of classical PSO by improving convergence velocity, solution quality, and robustness in high-dimensional, multimodal search landscapes. The proposed algorithm incorporates Lévy Flight (LF) to facilitate adaptive, long-distance position updates for superior global exploration, and integrates a Doppler Effect (DE)-inspired mechanism to dynamically modulate inertia weights, enhancing the balance between exploration and exploitation throughout the search process. The algorithm’s efficacy is evaluated through a comprehensive set of six constrained optimisation benchmark functions and a mechanical design problem (the speed reducer design) widely used in engineering optimisation research. PSO-LFDE’s performance is benchmarked against 13 state-of-the-art algorithms including α-Simplex, ASCHEA, CRGAL, CULDEV, DSS-MDEV2, DEVLC, HEA-ACT, ISR, MBA, PSO-DEV, SAPF, SMES, and SR. For the speed reducer optimisation, comparative analysis is performed against DSS-MDEV2, DEVLC, HEA-ACT, MBA, PSO-DEV, mbDEV, and the Society and Civilisation Algorithm (SC). Simulations are implemented in MATLAB R2017a, with real-time validation via Arduino-based PID-controlled DC motor hardware to assess practical feasibility. Results demonstrate that PSO-LFDE consistently achieves superior best and mean fitness values, significantly reducing standard deviation across all benchmark functions, indicating enhanced stability. In the speed reducer design task, PSO-LFDE yields the most optimal parameter configuration. Mean Absolute Error (MAE) analysis further validates the algorithm’s accuracy, with PSO-LFDE ranking first among 14 algorithms on benchmark functions (MAE = -893.4400), and first among 8 on engineering design problems (MAE = -631.0300), surpassing algorithms such as DSS-MDEV2 and HEA-ACT. In control system tuning, PSO-LFDE enhances the dynamic performance of a DC motor, achieving up to 86.78% faster settling time, 70.60% reduction in peak time, and lower steady-state error compared to conventional PSO. Across multiple set-point inputs for both position (10, 30, 60 cm) and speed (5, 10, 15 rpm) control, PSO-LFDE demonstrates consistently improved rise time, settling time, and overshoot reduction. The combination of swarm intelligence and physics-inspired strategies in PSO-LFDE establishes a promising framework for high-performance optimisation in constrained environments, with significant implications for engineering design, intelligent control, and real-time embedded systems. 
<br/><br/><label>ISBN: </label>THE0010180 (Local) 
<br/><br/><label>Subjects--Corporate Names: </label><br/>Faculty of Manufacturing & Mechatronics Engineering Technology--Dissertations
<br/><br/><label>Subjects--Topical Terms: </label><br/>Universities and colleges--Dissertations<br/>Thesis--Dissertations