The present invention relates to image satellite planning, and more particularly to a method for allowing a deeper search for high value targets in a time-limited planning environment. In an exemplary embodiment, a method of computing an ordered subset of targets includes using an approximation for the time needed for the satellite to re-orient to a new target, rather than calculating each maneuver time between targets. By approximating the maneuver time rather than calculating it, the calculation time is reduced. Each iteration through the traveling salesman problem takes less time, and more iterations can be accomplished between imaging windows. The iterative process can search deeper into the traveling salesman problem to find a better solution.