On finding an ear decomposition of an undirected graph distributively

نویسنده

  • Yung H. Tsin
چکیده

An ear decomposition of a connected undirected graph G= (V ,E) is a partition of the edge (link) set E into a collection of edge-disjoint paths P0,P1, . . . ,P|E|−|V |+1, called ears, such that P0 is a link, P0 ∪P1 is a closed path (i.e., a cycle), and for each i , 2 i |E| − |V | + 1, Pi is a path of which each terminating node belongs to some Pj , j < i , and no internal node belongs to any Pj , j < i . It has been used as a means in solving graph problems such as st-numbering, planarity testing, and graph connectivity in the sequential and parallel settings. Recently, Kazmierczak and Radhakrishnan [2] presented a distributed algorithm for constructing an ear decomposition of a 2-edge-connected (i.e., connected and bridgeless) undirected graph. They claimed that their algorithm has an O(m) time and message complexity where m is the number of links in the graph (Theorem 4 of [2]2). We shall show that the message complexity of their algorithm is actually (mn), where n is the number of nodes in the graph. Unfortunately, there is no easy fix for the error without completely changing the algorithm. This is owing to the fact that the algorithm uses a top-down approach to determine the ears which requires corrections to be made repeatedly over some part of the graphs resulting in substantial overhead, and a depth-first search method that would not shift from the current node to another node until the current node has received an acknowledgment for every message it sent out which results in long delay. We then outline a distributed algorithm that requires 2n− 2 time and 4m− n+ 1 messages. As these time and message complexity are the same as those of the best known distributed depth-first search algorithms [3,5], this is the best result one could ever achieve with a depth-first search based distributed algorithm [3]. Besides being more time and message efficient, our algorithm also uses a distributed model weaker then that of Kazmierczak et al. in that it does not require the messages sent along the

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

The Main Eigenvalues of the Undirected Power Graph of a Group

The undirected power graph of a finite group $G$, $P(G)$, is a graph with the group elements of $G$ as vertices and two vertices are adjacent if and only if one of them is a power of the other. Let $A$ be an adjacency matrix of $P(G)$. An eigenvalue $lambda$ of $A$ is a main eigenvalue if the eigenspace $epsilon(lambda)$ has an eigenvector $X$ such that $X^{t}jjneq 0$, where $jj$ is the all-one...

متن کامل

Parallel Open Ear Decomposition with Applications to Graph Biconnectivity and Triconnectivity

This report deals with a parallel algorithmic technique that has proved to be very useful in the design of eecient parallel algorithms for several problems on undirected graphs. We describe this method for searching undirected graphs, called \open ear decomposition", and we relate this decomposition to graph biconnectivity. We present an eecient parallel algorithm for nding this decomposition a...

متن کامل

Chapter in SYNTHESIS OF PARALLEL ALGORITHMS

This report deals with a parallel algorithmic technique that has proved to be very useful in the design of e cient parallel algorithms for several problems on undirected graphs. We describe this method for searching undirected graphs, called \open ear decomposition", and we relate this decomposition to graph biconnectivity. We present an e cient parallel algorithm for nding this decomposition a...

متن کامل

Optimal Randomized EREW PRAM Algorithms for Finding Spanning Forests

We present the rst randomized O(log n) time and O(m+n) work EREW PRAM algorithm for nding a spanning forest of an undirected graph G = (V; E) with n vertices and m edges. Our algorithm is optimal with respect to time, work and space. As a consequence we get optimal randomized EREW PRAM algorithms for other basic connectivity problems such as nding a bipartite partition, nding bridges and biconn...

متن کامل

Optimal Randomized Erew Pram Algorithms for Nding Spanning Forests and for Other Basic Graph Connectivity Problems

We present the rst randomized O(logn) time and O(m + n) work EREW PRAM algorithm for nding a spanning forest of an undirected graph G = (V; E) with n vertices and m edges. Our algorithm is optimal with respect to time, work and space. As a consequence we get optimal randomized EREW PRAM algorithms for other basic connectivity problems such as nding a bipartite partition, nding bridges and bicon...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:
  • Inf. Process. Lett.

دوره 91  شماره 

صفحات  -

تاریخ انتشار 2004