note_on_counting_eulerian_circuits研究生英文论文.pdfVIP

Note on Counting Eulerian Circuits Graham R. Brightwell ∗ Peter Winkler † April 8, 2006 4 0 0 2 y Abstract a M We show that the problem of counting the number of Eulerian circuits in an undirected graph is complete for the class #P. 8 1 1 1 Introduction v 7 6 Every basic text in graph theory contains the story of Euler and the K¨onigsberg bridges, 0 5 together with the theorem that guarantees the existence of a circuit traversing every edge of a 0 4 graph exactly once, if and only if the graph is connected and all vertices have even degree. 0 / Later in the text one might find the “Matrix Tree Theorem”, which provides an efficient C C algorithm for counting the number of spanning trees of a graph; and an application of this . s and the so-called “BEST” Theorem (see below) to count the number of Eulerian circuits in a c : v directed graph. But what about counting Eulerian circuits in an undirected graph? i X This problem is clearly in the class #P (introduced by Valiant [11] in the 1970s), since r a it is easy to check whether a candidate circuit traverses each edge once. Since there is no known efficient way to count Eulerian circuits, it is natural to suspect that the problem is #P- complete, and thus presumably very difficult—especially in view of Toda’s result [9], which implies that one call to a #P oracle suffices to solve any problem in the polynomial hierarchy in deterministic polynomial time. Researchers have shown a myriad of graphical counting problems to be #P