Highly dynamic information environments that support military applications should be able to satisfy increasingly stringent demands on end-to-end QoS. As part of the DARPA AICE (Agile Information Control Environment) program, we have developed a middleware framework for admission control and resource allocation, for maximization of the end-to-end QoS for communication channels. Called Adaptive Information Control (AIC), the framework includes a novel architecture for the control environment, and rich models for flow characterization and resource constraints. We describes novel algorithms, based on strict priority, for resource allocation that maximize the overall mission objectives. A concept of 'bottleneck priority' is introduced which simplifies the computational effort. The overall architecture is described along with various resource allocation algorithms. Different test cases demonstrate the utility of such optimal resource allocation mechanisms for the success of the overall mission.