On the Convergence and Diversity of Pareto Fronts Using Swarm Intelligence Metaheuristics for Constrained Search Space

This article reviews existing constraint-handling techniques then presents a new design for Swarm Intelligence Metaheuristics (SIM) to deal with constrained multi-objective optimization problems (CMOPs). This new design aims to investigate potential effects of leader concepts that characterize the dynamic of SIM in the hope to help the population to reach Pareto optimal solutions in a constrained search space. The new leader-based constraint handling mechanism is incorporated in Constrained Multi-Objective Cuckoo Search (C-MOCS) and Constrained Multi-Objective Particle Swarm Optimization (C-MOPSO) as specific instances of a more general class of SIMs. The experimental results are carried out using a set of six well-known test functions and two performance metrics. The convergence and diversity of C-MOCS and C-MOPSO are analysed and compared to the well-known Multi-Objective Evolutionary Algorithm (MOEA) NSGA-II and discussed based on the obtained results.