This paper introduces a two-tier framework for energy management in mobile devices. The first tier is a static tier based on the Analytical Hierarchy Process that highlights a hierarchy of factors for energy management decisions. This tier can be used at the architectural level, where decisions are not dependent on the environment. The second tier is a dynamic tier making use of neural networks to evaluate environmental factors dynamically and adjust accordingly the energy profile of the mobile device's components and the computational priorities. The framework allows greater flexibility than a single energy policy and improves on user-controlled policy switching because it can monitor environmental factors in real time.