Convergence and Diversity Measurement for Vector Evaluated Particle Swarm Optimization Based on ZDT Test Problems

Vector Evaluated Particle Swarm Optimization (VEPSO) has been successfully applied to various applications. However, the VEPSO actual performance is still uncertain. Hence, this paper will evaluate the VEPSO performance in term of convergence and diversity ability using Generalized Distance and Spread measurement respectively. Simulation with ZDT benchmark test problems show VEPSO is weak in solving non-convex, non-uniformity search space and low solution density near Pareto optimal front problems. Besides, VEPSO is very weak in multi modality problems because PSO weakness in facing multiple local Pareto optimal fronts problems. Lastly, VEPSO has weak diversity ability due to no diversity control mechanism in searching the solutions.