Description:
TECHNOLOGY AREA(S): Weapons
OBJECTIVE: Develop a simulation architecture using the precepts of evolutionary algorithms to optimize and adapt a multi-disciplinary autonomous munition model. The munition model would provide an evolving design baseline for research and development as knowledge of the target environment changes.
DESCRIPTION: Evolutionary algorithms (EAs) provide a means to optimize designs that are highly complex and involve a large multi-disciplinary parameter space. This topic will use EAs to optimize guided weapons, as represented by closed-loop simulations that capture performance changes due to design variations in guidance, navigation, control, airframe, propulsion, mass properties, warhead, sensor parameters, etc.
PHASE I: Perform initial development of an evolutionary architecture, selecting key parameters to evolve, identifying fitness criteria, exploring parent selection and child generation techniques, and establishing bounds on random variations. Establish a simulation baseline, exploring 3DOF vs 6DOF models and identifying the generality and detail of component models to evolve. Identify an engagement scenario and bound target variations for an initial demonstration.
PHASE II: Expand the free parameter space to encompass a wider range of design variations. Incorporate airframe and mass property changes along with a means to estimate aerodynamic properties to feed airframe truth models (Missile Datcom?). Migrate to a standard 6DOF weapon architecture (AFRL UPBEAT?). Design and procure a hardware architecture for long term simulation hosting and data archiving. Establish a protocol for runtime updating of target sets and variations (AFSIM?). Provide a runtime visualization capability. Investigate transition of specific “living” weapon concepts.
PHASE III: Create and implement new lines of weapon concepts for alternative missions, based on range, launch platform, and target set. Investigate air to air, surface to air, close air support, long range strike, and others based on customer demand. Investigate application of technology to higher level simulation environments at the mission and campaign level.
REFERENCES:
1: Gareth Jones. Genetic and evolutionary algorithms. In Paul von Ragu´e Schleyer and Johnny Gasteiger, editors, Encyclopedia of Computational Chemistry, volume III - Databases and Expert Systems, chapter 2 John Wiley & Sons, Ltd., September 199 ISBN: 978-0-47196-588-
2: Thomas Back, ‘Evolutionary Algorithms in Theory and Practice’, Oxford University Press, Oxford, 199 ISBN: 9780195099713
3: D. E. Goldberg, ‘Genetic Algorithms in Search Optimization and Machine Learning’, Addison-Wesley, Reading, MA, 198
KEYWORDS: Evolutionary Algorithms, Genetic Algorithms, Optimization, Weapon, Munitions, Guidance And Control, Closed Loop
CONTACT(S):
Rhoe A. Thompson
(850) 883-0867
rhoe.thompson@us.af.mil
