抄録
This letter studies a nesting discrete particle swarm optimizer for multi-solution problems. The algorithm operates in discrete search space and consists of two stages. The first stage is global search in rough lattice points for constructing local sub-regions each of which includes one target solution. The second stage is local search where the algorithm operates in parallel in fine lattice points of local subspaces and tires to find all the approximate solutions within a criterion. We then propose an application to finding multiple fixed points in nonlinear dynamical systems and investigate the algorithm efficiency.
