Labeling a type as optional is one purpose of nullable item/types. But that is really open to the design. but is see the point: If the designer of the system knows the information may not be available in some cases, label it as optional and then use nulls.
If the data item is part of a larger enterprise data dictionary, perhaps this null check could be done outside the DMN. For instance, I've seen Data Quality services in enterprise systems that would execute this type of validation logic (null, mask,length, range, enum, etc). outside of the system that runs the business logic.
making a decision "based on the predetermined required info" vs "based on the information available" are two viable approaches. Often times, the best decision needs to be made, acknowledging that some information may not be available at the time of the decision, thus the decision moves forward with the info it was provided. For instance, in human lung-organ donors/donation, frequently a bronchoscope video is performed to provide additional information of the health of the organ. The results of the bronchoscope are valuable in defining the suitability of the organ, however a bronchscope video is not always performed/provided, and the suitability decision needs to be made regardless.