Resource-Bound Swarm Controller: Velocity-Clamped, Boundary-Safe PSO for Embedded and Real-Time Optimization

• Monisha Rengaraj

  Author

This paper presents a processor-friendly swarm optimization controller that combines velocity clamping with boundary-safe particle penalization and a tailored inertial-weight schedule to deliver fast, predictable convergence under strict compute and memory constraints common to embedded and real-time systems. Recent work has explored PSO adaptations for resource-constrained and embedded environments, emphasizing computational efficiency and bounded execution behavior. The method enforces feasibility at every update via bounded velocity [V_min, V_max] and a zero-velocity reflection when proposed steps exceed search limits, while a custom inertial profile avoids stagnation and accelerates descent compared to constant or linearly decreasing weights. A standardized evaluation on seven 30-dimensional benchmark functions with termination FE = 5000 x D  quantifies mean/variance improvements and convergence speedups against multiple PSO baselines, highlighting robustness across unimodal and multimodal objectives. Design-for-deployment guidance is provided for microcontrollers and SoCs, emphasizing fixed-cost bound checks, parameter ranges for stability, and portability of the update rules, which together enable efficient firmware or accelerator integration without sacrificing solution quality.

• PDF

Copyright (c) 2026 International Journal of Intelligent Systems and Data Science

This work is licensed under a Creative Commons Attribution 4.0 International License.