Autonomous Negotiating Teams (ANTS)

(supported by DARPA, via subcontract from Northeastern Univ., BAE-Sanders (formerly Lockheed-Sanders), and SRI)

Stability Analysis of Distributed Scan Schedulers

K. Passino


The ANTS program focuses on developing theoretical foundations and practical technical solutions for a new class of real-time distributed resource allocation systems. Autonomous negotiation of assignment and efficient use of resources in highly decentralized environments, making maximum use of local information, such as weapons to tasks, with moving weapons/targets is being considered. The goal is to build systems that exploit available distributed, networked computational resources (hardware, operating systems, and communication).

The OSU focus is on "distributed dynamic scan scheduling" (DDSS). First, we have considered dynamic scan scheduling (DSS) (rather than the current fixed scan scheduling) for a single aircraft-based "receiver" that seeks to allocate its resources to maintain an accurate assessment of an environment with many "emitters." For this we have developed a wide variety of DSSs, shown how to design such strategies to optimize performance measures (relating to how long each emitter is ignored by the receiver), and conducted stability analysis to show that several DSSs provide stable operation if a certain "capacity condition" is satisfied (i.e., if the receiver's workload is not too high). The plot below shows (on top) the pattern of emitters that a receiver chooses to focus on, and the bottom shows the maximum time that any one emitter is ignored.

.

Building on our work with DSS we also studied the design and stability analysis of DDSS, and did simulations to show its potential advantages.

You can obtain a report by clicking here, or see the main project web site by clicking here.