This paper describes algorithms for automatic scheduling and deconfliction of satellite communication requests on highly constrained antenna resources. In addition to initially honoring the constraints included in the communication requests, the system also resolves remaining conflicts by suggesting "bending the rules" to resolve the conflicts based on past precedents retrieved from a case base. A prototype implementation is described along with the results from executing it on realistic data

more »

... ts. Future work is also presented. I. Problem Description Scheduling and optimizing satellite communication resources is an enormously complex task. There are many constraints to be considered. The satellite must have LOS and be in-range of the antenna that it communicates with during the entire window for which communication is being requested. Different stations have antennas with different capabilities and have different support equipment so different satellites will have different sets of ground stations they can communicate with. Most satellites have requirements to have a certain number of communication events per day with maximum and minimum allowed time separations between events. Especially for Low Earth Orbit (LEO) satellites, often the required communication event time period is the entire, or a very large fraction of, each calculated view period window. Additionally some amount of time is required between communication events to change the configuration, including repositioning the antenna. Communications resources are very expensive and have become very tight, making optimal use of those resources extremely important. There is a very strong goal of meeting ALL communications requests, regardless of priority. To accelerate satellite communication scheduling and deconfliction and to handle increased complexity of new resources, there is a need for a fully automated satellite communication resource planner which would receive communication requests and generate an optimized/deconflicted schedule that assigned resources for specific time windows to best meet those requests, considering the applicable constraints. The scheduling, optimization, and deconfliction process could be fully automated or performed interactively with human involvement, as desired by the user. Currently, a first pass schedule is generated automatically, which leaves conflicts to be resolved by the human scheduler. There are two main categories of conflicts -those that can be resolved while satisfying all the constraints in the requests, presumably by shuffling other resource assignments in allowable ways (including reducing communication events to their minimum allowed time periods from their preferred values) and those that cannot. Resolving the former is relatively straightforward but does involve the use of good heuristic algorithms to produce as good a result as possible. Resolving the latter requires actually deciding what requests should have their constraints relaxed somewhat. This is currently done by human schedulers based on past experience and precedent. An automated deconflictor must also follow past precedent by having a case base containing previous requests to the satellite communication network clients to relax constraints (called "suggestions"), along with whether the suggestion was accepted or denied. Then, when trying to resolve a current conflict, the system can first calculate which constraint relaxation suggestion would resolve the conflict and check the case base to determine if similar