Multi-robot systems are becoming increasingly popular in warehouses and factories, since they potentially enable the development of more versatile and robust systems than single robots. Multiple robots allow performing complex tasks with greater efficiency. However, this leads to increased complexity in planning and dispatching actions to robots. In this letter, we tackle such complexity using a hierarchical planning framework: the task is first planned at an abstract level and then refined by local motion planning. We propose a framework based on a state-transition system formalism that abstracts the problem by removing unnecessary details and, hence, considerably reduces planning space complexity. Forward search from an initial state allows the robot to find a sequence of actions to accomplish the assigned task. These actions can be planned at a lower level employing any motion planning technique available in the literature. The proposed method is validated through experiments in several operating conditions and scenarios.