Cross-Layer Utility-Based System Optimization

Multilevel fusion systems need provisions to optimally schedule scarce processing and communication resources. To this end, we explore the idea of using utility-based metrics to optimize the run-time operation of a computation and communication constrained multilevel system, including automatic decision support measures such as course of action planning.In a simplified case study, we extend a processing chain with local utility-based run-time feedback mechanisms. This allows the processes in the distributed chain to improve their local interactions only. Simulation results show an increased performance of utility-based optimization over constant rate or random techniques.