Computing Shortest Paths in Series-Parallel Graphs in Logarithmic Space

Series-parallel graphs, which are built by repeatedly applying series or parallel composition operations to paths, play an important role in computer science as they model the flow of information in many types of programs. For directed series-parallel graphs, we study the problem of finding a shortest path between two given vertices. Our main result is that we can find such a path in logarithmic space, which shows that the distance problem for series-parallel graphs is L-complete. Previously, it was known that one can compute some path in logarithmic space; but for other graph types, like undirected graphs or tournament graphs, constructing some path between given vertices is possible in logarithmic space while constructing a shortest path is NL-complete.