← Back to FPA docs

teeline · algorithms/fpa

Flower Pollination Algorithm

A population of candidate tours (flowers) evolves each step: with probability p a flower makes a long-range Lévy flight toward the global best; otherwise it cross-pollinates with two random flowers nearby. Drag the sliders to see how p and population size shape search.

01234567891011
Flowers
02169active flower
12320quality 41%
22122quality 78%
32057quality 90%
42230quality 58%
52003global best
62145quality 74%
72539quality 0%
● gbest tour● active flower● population● city▌ active flower▌ global best
Press Step or Run to begin
■ global Lévy step  ■ local ε step
step size
—
iteration
0
gbest distance
2003.0
global steps
0
local steps
0
~80% global / 20% local
Balanced population size
cities: 12flowers: 8p: 0.80β (Lévy): 1.5