Given a large but rather sparse graph (of a road network, say), the task is to efficiently generate all existing minimal multiway edge-cuts up to a certain number of edges. The student will implement and experimentally evaluate a new "circuit-cocircuit" algorithm (proposed by the advisor, unpublished) for this problem. Special attention will be paid to a practically fast implementation, and to so called canonical generation of the cuts (i.e., without repetition, achieved by observing certain unique canonical order for all the cuts).