Recent advancements in mobile computing and communication technologies has given rise to development of mobile ad hoc computational grids. The mobile ad hoc computational grids provide a software infrastructure to share computing resources in mobile ad hoc environments. The adaptation of computational grids in mobile ad hoc environments is not straightforward, and presents many research challenges due to node mobility and infrastructure-less network environment. In order to improve utilization of shared computing resources and application performance, one of the key services is a resource allocation. In this paper, we propose a centralized resource allocation scheme that allocates dependent tasks among nodes in mobile ad hoc computational grids. First, we classify tasks and dependencies, and then exploit them to reduce communication cost among interdependent tasks, and thus application completion time. The effectiveness of proposed scheme is demonstrated through simulations.