This paper addresses the problem of energy efficient resource allocation to interdependent tasks in mobile ad hoc computational grids. The interdependent tasks may exchange large quantity of data and consume the energy proportional to amount of data exchanged among the tasks. If tasks wouldn't be allocated effectively, then there can be a significant increase in the energy consumption and communication cost. The increased energy consumption limit the life time of a node and may result in a power failure which not only affects the task executing on a node but also affects other tasks in various ways. In addition, the mobile nodes within a grid are battery driven and thus have a limited power that should be utilized effectively. In this paper, we propose a hybrid power-based resource allocation scheme to reduce the energy consumption and communication cost among the interdependent tasks. The basic idea is to exploit the dependency and task types and allocate interdependent tasks to nodes accessible at a minimum transmission power level. We also propose a power-based algorithm to find a group of closest nodes to allocate a set of interdependent tasks. Compared to the traditional algorithms, the complexity of proposed algorithm depends on the number of transmission power levels rather than the number of nodes within a grid.