GRP1. A recursive fusion operator for the transferable belief model

Generally, there are problems with any form of recursive fusion based on belief functions. An open world is often required but the empty set can become greedy where, over time, all of the mass will become assigned to it where conjunctive combination is used. With disjunctive combination, all of the mass will move toward the ignorant set over time. In real world applications often the problem will require an open world but due to these limitations it is forced into a closed world solution. GRP1 works iteratively in an open world in a temporally conscious fashion, allowing more recent measurements to have more of an impact. This approach makes it ideal for fusing and classifying streaming data.