What order of magnitude are you going to be working on? Thousands of elements? Millions?
Boost is linked to your code like any other library, you might want to have a look at: http://www.boost.org/doc/libs/1_39_0/more/getting_started/unix-variants.html
Other than this, you might want to have a look at Networkx, or iGraph (for Python) for rapid prototyping purposes, if you want to quickly test an algorithm (https://networkx.github.io/ and http://igraph.org/redirect.html).
I have used Networkx with datasets of a few hundred thousand nodes without any problems. Other than this, graphviz can be used to visualise graphs with millions of nodes but you are going to need a machine with a lot of memory (>6GB). Typical rendering times for such tasks might go well over 2hrs (http://www.graphviz.org/ and http://yifanhu.net/GALLERY/GRAPHS/index.html).