← Back to PSO docs

teeline · algorithms/pso

Particle Swarm Optimisation

A swarm of tour-particles updates each epoch. Each particle's velocity is a list of city-swap moves built from three components: ω keeps momentum from the previous epoch, C₁ pulls toward the particle's own personal best, and C₂ pulls toward the global best. Inertia ω decays over time, shifting the swarm from exploration toward exploitation.

01234567891011
Particles
02406
dist 2406
12097
dist 2097
22351
dist 2351
31993
dist 1993
42100
dist 2100
51802
gbest
gbest tour best particle swarm city
Press Step or Run to begin
ω (inertia) = 0.900
High — broad exploration
epoch
0
gbest dist
1801.5
avg dist
2125
v_max
5 swaps
Balanced swarm size
cities: 12particles: 6ω: 0.4→0.9C₁=C₂=1.5v_max: 5