matlab - Particle Swarm Optimization: what means gBest value? -

i designing particle swarm optimization simulator matlab , have doubt gbest local position.

as understood reading algorithm, gbest value identify global best value has ever discovered particles. however, saw in pretty bearable implementation developer considered gbest best value encountered particles in current generation.

have misunderstood pso algorithm basics? developer wrong in implementation or, however, different approach?

thanks in advance, víctor.

there many ways fine tune pso , classical algorithm of pso uses gbest swarm best known position, therefore not current iteration only. in algorithm have implement memory of sort in order calculate delta used in velocity update:

note called pbest particle best known position , sbest swarm best known position (your gbest). c1 , c2 confidence levels in particle , in swarm.