A stochastic optimization framework for resource management and course of action analysis

Abstract

We consider a course of action (COA) problem domain with the support of information fusion technologies. We assume that level 1 and 2 fusion technologies have provided a set of suspicious/abnormal situations currently occurring and associated dasiaobjects of interestpsila participating in these situations, along with likelihoods of participation. We further assume that there are limited resources available to gather additional information on these objects of interest. COA analysis and resource management (RM) prescribe placement of resources to accomplish desired information gathering tasks. Thus the goal of COA/RM is ldquoputting the right resource in the right place and time to perform the appropriate information gathering task on the appropriate objects of interest.rdquo Our research focuses on formulating the COA/RM problem from a stochastic optimization framework, resulting in an NP-hard problem. An efficient heuristic is developed and demonstrated on an illustrative scenario from maritime domain awareness.