A Search Method Combining Average Distance of Particle Swarm and Crowding Distance of Particles

It is popular to constantly improve existing optimization algorithms by embedding intelligent operators. In this paper, with particle swarm optimization algorithm (PSO) as the framework, the average distance of particle swarm is combined with the crowding distance of particles to provide conveniences for selecting the global optimal particle and deleting redundant particles in Pareto optimal set (PS). What's more, exponential weight and adaptive mutation are used to control the trade-off between global and local exploration. The proposed method has been evaluated by inverted generation distance (IGD) and compared with several typical multi-objective optimization algorithms. Experimental results show that the proposed method can improve the ability of searching for unknown solutions and also obtain Pareto front (PF) with good distribution on most of standard functions.