Multi-constraint Evolutionary Algorithm Based on Decomposition for the Synthesis of Concentric Ring Sparse Array

For the local convergence phenomenon in the multiobjective evolutionary algorithm based on decomposition (MOEA/D) when applied to multiconstraint sparse array optimization, a multiconstraint evolutionary algorithm based on decomposition (MCEA/D) is proposed in this communication. Unlike MOEA/D, which constructs subproblems through weighted aggregation of multiple objectives, MCEA/D converts objective functions into constraints while retaining a single objective function, generating subproblems via constraint value space decomposition. Consequently, MCEA/D performs optimization within different constraint subspaces, with each subproblem operating in its own search space. This novel approach effectively mitigates the diversity loss and premature convergence issues of MOEA/D caused by the shared search space among subproblems. The results of uniformly excited concentric ring sparse array (CRSA) optimization demonstrate that MCEA/D can successfully identify feasible Pareto fronts (PFs), outperforming MOEA/D algorithm in this capability. Electromagnetic simulations validate the results.