Per the MagicGrid methodology, is it acceptable to rearrange and repartition the problem domain conceptual subsystems in the solution domain logical subsystems?

For example, a radar can have a conceptual antenna subsystem, a conceptual transmitter substem, and a conceptual receiver subsystem. In practice the antenna module/subsystem can include parts of the transmitter conceptual subsystem and parts of the receiver conceptual subsystem in addition to the antenna itself. Another solution can choose to partition things differently. 

How would the relationships show the partitioning and repackaging of the conceptual subsystems. In the MagicGrid Book Of Knowledge, there  is one-to-one mapping between the problem domain conceptual systems and the solution domain conceptual subsystems so I'm confused.

Thanks