Skip to main content

> ML_ALGORITHM // PARTICLE-SWARM-OPTIMIZATION-PSO_v1.0

Particle Swarm Optimization (PSO)

Computational intelligence technique that optimizes candidate solutions by having a swarm of particles explore the search space driven by social and cognitive velocities.

Swarm Intelligenceevolutionary-searchhigh-intrinsicsmall (<1k)
Back to All Algorithms
Computational Complexity
Training Complexity:O(iterations * particles * d)
Inference Complexity:O(1)
Hardware Profile
CPU Friendly:Yes
Requires GPU:No
Memory Footprint:low
Interpretability & Data
Interpretability Tier:high-intrinsic
Training Data Needs:small (<1k)

Interpretability Assessment

Particle trajectory kinematics can be visualized as physical trajectories in feature space.

Suitable Tasks & Supported Modalities

Suitable Tasks:
continuous blackbox optimizationfeature selection
Supported Modalities:
tabular

Implementing Libraries

pyswarms
deap

Foundational Literature

Common Pitfalls & Warnings
  • Velocity explosion without inertia weight decay or velocity clamping constants (v_max)