A bi-objective multi-swarm particle swarm optimization algorithm for the frequency assignment problem

Nearly all previous works on the frequency assignment problem (FAP) have modeled it as an optimization problem with a single optimization objective, which is either the total interference or number of violations. Solutions obtained by solving these single-objective optimization problems can only meet one of the two requirements. A bi-objective optimization problem (BOOP) has been proposed to model the FAP, and a multi-swarm particle swarm optimization (MS-PSO) algorithm has been proposed to solve the BOOP. In order to avoid the MS-PSO algorithm converging to the local minimum, a condition which the optimal solution satisfies has been proposed and employed to determine the direction of the search. Simulations show that the solutions of the proposed bi-objective optimization model can made trade-off between the two requirements, and the optimal condition guided MS-PSO algorithm can input better solution than that conventional MS-PSO algorithm.