In this research a scheduling algorithm is proposed for transmission of information via a satellite to receivers on the ground. The satellite contains a number of antenna-beam areas, which are serviced by a smaller number of agilely switched burst transmitters. The burst times are constant and synchronized across all transmitters. The problem is to develop a near optimal schedule for the transmission of information based upon an evolving beam topology (determined by the orientation of the satellite) along with stringent frequency bandwidth overlap constraints, i.e. messages sent with the same frequency must be sufficiently scattered across the beam areas. Thus, this scheduling problem can be interpreted as a combination of a traditional scheduling problem as well as a map coloring problem. We dynamically form transmitter conflict graphs and exploit fast clustering algorithms to efficiently solve this scheduling problem. Our greedy optimization framework offers near optimal schedule and can be employed on the limited on-board satellite computing hardware given a short time horizon.