Virtual network embedding is a process that allows for the creation of multiple virtual networks, each specifically designed to meet the unique requirements of various services, on a shared physical network infrastructure. 5G network virtualization requires an efficient embedding algorithm to maximize the resource utilization of the substrate network. In this paper, we propose a hierarchical cooperative multi-agent reinforcement learning scheme to provide a virtual network embedding solution for effective network resource utilization, maximum revenue, and minimum cost. The proposed scheme applies 1) hierarchical reinforcement learning for efficient exploration by dividing the problem into sub-steps, and 2) multiagent-based cooperative reinforcement learning to improve algorithm performance through the cooperation of multiple agents. The simulation results show that the proposed scheme outperforms the comparative schemes in terms of revenue, cost, revenue-to-cost ratio, and acceptance ratio.