Task assignment is paramount for the efficiency of multi-robot missions. In the context of multi-robot exploration of unknown environments, tasks can be identified as targets to reach in order to expand knowledge of the environment. These targets can be frontiers between known reachable area and unknown areas or a best view configuration to observe each frontiers. Once these tasks are identified, they are assigned to robots. Standard approaches work in a two step process where the costs of reaching targets are computed before running a task allocation algorithm. In this article, we combine these two steps to propose an efficient task allocation algorithm. This algorithm can take into account different assignment criteria such as the rank of a robot in a list of robots ordered by cost.