A New Hybrid Particle Swarm Optimization and Evolutionary Algorithm

Particle swarm optimization (PSO) has proved fast convergence in many optimization problems but still has the main drawback falling in a local minimum. This paper presents a new Hybrid Particle Swarm Optimization and Evolutionary algorithm (HPSO-E) to solve this problem by introducing a new population of children particles obtained by applying a mutation and crossover operators taken from the evolutionary algorithm. In this way, we connect the best properties of the algorithms: fast convergence of the PSO and ability to global search introduced by the evolutionary algorithm. The novel hybrid algorithm shows sufficient convergence for unimodal benchmark function and excellent convergence for selected hard multimodal benchmark functions.