Share Email Print

Proceedings Paper

An evolutionary algorithm technique for intelligence, surveillance, and reconnaissance plan optimization
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

To support an Effects Based Approach to Operations (EBAO), Intelligence, Surveillance, and Reconnaissance (ISR) planners must optimize collection plans within an evolving battlespace. A need exists for a decision support tool that allows ISR planners to rapidly generate and rehearse high-performing ISR plans that balance multiple objectives and constraints to address dynamic collection requirements for assessment. To meet this need we have designed an evolutionary algorithm (EA)-based "Integrated ISR Plan Analysis and Rehearsal System" (I2PARS) to support Effects-based Assessment (EBA). I2PARS supports ISR mission planning and dynamic replanning to coordinate assets and optimize their routes, allocation and tasking. It uses an evolutionary algorithm to address the large parametric space of route-finding problems which is sometimes discontinuous in the ISR domain because of conflicting objectives such as minimizing asset utilization yet maximizing ISR coverage. EAs are uniquely suited for generating solutions in dynamic environments and also allow user feedback. They are therefore ideal for "streaming optimization" and dynamic replanning of ISR mission plans. I2PARS uses the Non-dominated Sorting Genetic Algorithm (NSGA-II) to automatically generate a diverse set of high performing collection plans given multiple objectives, constraints, and assets. Intended end users of I2PARS include ISR planners in the Combined Air Operations Centers and Joint Intelligence Centers. Here we show the feasibility of applying the NSGA-II algorithm and EAs in general to the ISR planning domain. Unique genetic representations and operators for optimization within the ISR domain are presented along with multi-objective optimization criteria for ISR planning. Promising results of the I2PARS architecture design, early software prototype, and limited domain testing of the new algorithm are discussed. We also present plans for future research and development, as well as technology transition goals.

Paper Details

Date Published: 2 May 2008
PDF: 15 pages
Proc. SPIE 6964, Evolutionary and Bio-Inspired Computation: Theory and Applications II, 696407 (2 May 2008); doi: 10.1117/12.784272
Show Author Affiliations
John T. Langton, Charles River Analytics, Inc. (United States)
Joseph A. Caroli, Air Force Research Lab. (United States)
Brad Rosenberg, Charles River Analytics, Inc. (United States)

Published in SPIE Proceedings Vol. 6964:
Evolutionary and Bio-Inspired Computation: Theory and Applications II
Misty Blowers; Alex F. Sisti, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?