Abstract: A novel evolutionary algorithm is presented in order to improve the efficiency of pattern synthesis of antenna arrays.In the algorithm,the optimized variable vectors are simulated as members of a team.After the team members are divided into the elite and plain groups,the epitomes of both groups are determined,the manipulations of learning and exploration are properly defined,and the member renewal rules are reasonably established.The members' division of action can emerge during the search procedure,which makes the algorithm possesses the abilities of global,local and directional search.Numerical results verify that the new algorithm has the properties of simple implementation,fast convergence,low computation cost and easy parameter selection.Examples of amplitude-only,phase-only and amplitude-phase excitations of antenna arrays are given.The algorithm can be employed to solve similar optimization problems.