Using Evolutionary Programming to Optimize the Allocation of Surveillance Assets

Abstract

An intelligent surveillance planning system must allocate available resources to optimize data collection with respect to a variety of operational requirements. In addition, these requirements often vary temporally (i.e., targets of interest move, priorities change, etc.), requiring dynamic reoptimization on-the-fly. Allocation of surveillance resources has typically been accomplished either by human planners, (for small problems of very limited complexity) or by deterministic methods (typically producing suboptimal solutions which are incapable of adapting to dynamic changes in the environment). The method presented here solves these problems by using evolutionary programming to optimize the simultaneous and coordinated scheduling of multiple surveillance assets. The problem of allocating unmanned aerial vehicles (UAVs) to acquire temporally variable, time-differential intelligence data is addressed. Imposition of realistic constraints ensures solution feasibility in real-world problems. This implementation can be modified to optimize solutions for a suite of different surveillance asset types, such as manned vehicles and satellites.