Share Email Print

Proceedings Paper

Neural network submodel as an abstraction tool: relating network performance to combat outcome
Author(s): Greg Jablunovsky; Clark Dorman; Paul S. Yaworsky
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Simulation of Command and Control (C2) networks has historically emphasized individual system performance with little architectural context or credible linkage to `bottom- line' measures of combat outcomes. Renewed interest in modeling C2 effects and relationships stems from emerging network intensive operational concepts. This demands improved methods to span the analytical hierarchy between C2 system performance models and theater-level models. Neural network technology offers a modeling approach that can abstract the essential behavior of higher resolution C2 models within a campaign simulation. The proposed methodology uses off-line learning of the relationships between network state and campaign-impacting performance of a complex C2 architecture and then approximation of that performance as a time-varying parameter in an aggregated simulation. Ultimately, this abstraction tool offers an increased fidelity of C2 system simulation that captures dynamic network dependencies within a campaign context.

Paper Details

Date Published: 23 June 2000
PDF: 9 pages
Proc. SPIE 4026, Enabling Technology for Simulation Science IV, (23 June 2000); doi: 10.1117/12.389371
Show Author Affiliations
Greg Jablunovsky, SM&A Corp. (United States)
Clark Dorman, SM&A Corp. (United States)
Paul S. Yaworsky, Air Force Research Lab. (United States)

Published in SPIE Proceedings Vol. 4026:
Enabling Technology for Simulation Science IV
Alex F. Sisti, Editor(s)

© SPIE. Terms of Use
Back to Top