A recently proposed metaheuristics called Wolf Search Algorithm (WSA) has demonstrated its efficacy for various hard-to-solve optimization problems. In this paper, an improved version of WSA namely Eidetic-WSA with a global memory structure (GMS) or just eWSA is presented. eWSA makes use of GMS for improving its search for the optimal fitness value by preventing mediocre visited places in the search space to be visited again in future iterations. Inherited from swarm intelligence, search agents in eWSA and the traditional WSA merge into an optimal solution although the agents behave and make decisions autonomously. Heuristic information gathered from collective memory of the swarm search agents is stored in GMS. The heuristics eventually leads to faster convergence and improved optimal fitness. The concept is similar to a hybrid metaheuristics based on WSA and Tabu Search. eWSA is tested with seven standard optimization functions rigorously. In particular, eWSA is compared with two state-of-the-art metaheuristics, Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO). eWSA shares some similarity with both approaches with respect to directed-random search. The similarity with ACO is, however, stronger as ACO uses pheromones as global information references that allow a balance between using previous knowledge and exploring new solutions. Under comparable experimental settings (identical population size and number of generations) eWSA is shown to outperform both ACO and PSO with statistical significance. When dedicating the same computation time, only ACO can be outperformed due to a comparably long run time per iteration of eWSA.