Self Aware Processing
Small Business Information
P.O. Box 6024, Sherman Oaks, CA, 91413
AbstractPredetermined, hard-wired processing architectures with static resource allocation and task distributions suffer from single points of failure, inefficient performance or failure in dynamic situations and have little to no run-time decision making ability concerning task redistribution and optimization. By using unique multi-agent system methodologies and leveraging core capabilities, Areté proposes to design a self-aware, agent-based system for development, parallelization, optimization and execution of code on distributed systems with the ability to adapt and re-optimize resources and tasks as situations dictate. Through continuous self-modeling and prediction, self-aware, self-healing systems can enhance current distributed system performance and allow for the deployment of systems in situations never before possible with current fault-tolerant techniques.
* information listed above is at the time of submission.