In this paper, we will discuss a map synthesis system which handles static
information (geographic objects) as well as dynamic information (traffic co
nditions, weather? etc.). In addition to geographic thesauruses used in the
previous systems, we will use co-existence relationships to improve the qu
ality of maps generation. The system is considered to be general purpose (n
ot restricted to car navigation nor travel maps) and can generate arbitrary
maps according to the user's specification. It is very difficult for a use
r to specify a query which corresponds to the required map, because map des
cription is not easy. The system should automatically generate missing info
rmation or find errors in the user specification. For the purpose we use ge
ographic domain thesauruses which contain aggregation and other geographic
relationships as well as conventional thesaurus hierarchy. In this paper, w
e will discuss to use co-existence relationships to enhance ability to sele
ct geographic objects automatically. Co-existence specifies relationships a
mong geographic objects which should appear in a map together although they
may not have geographic relationship by thesauruses. By utilizing co-exist
ence relationships, a user can acquire much more understandable maps.