The Metropolis algorithm for graph bisection

被引:67
作者
Jerrum, M
Sorkin, GB
机构
[1] Univ Edinburgh, Dept Comp Sci, Edinburgh EH9 3JZ, Midlothian, Scotland
[2] IBM Corp, Thomas J Watson Res Ctr, Dept Math Sci, Yorktown Hts, NY 10598 USA
关键词
D O I
10.1016/S0166-218X(97)00133-9
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
We resolve in the affirmative a question of Boppana and Bui: whether simulated annealing can, with high probability and in polynomial time, find the optimal bisection of a random graph in G(npr) when p - r = Theta(n(Delta-2)) for Delta less than or equal to 2. (The random graph model G(npr) specifies a "planted" bisection of density r, separating two n/2-vertex subsets of slightly higher density p.) We show that simulated "annealing" at an appropriate fixed temperature (i.e., the Metropolis algorithm) finds the unique smallest bisection in O(n(2+epsilon)) steps with very high probability, provided Delta > 11/6. (By using a slightly modified neighborhood structure, the number of steps can be reduced to O(n(1+epsilon)).) We leave open the question of whether annealing is effective for Delta in the range 3/2 < Delta less than or equal to 11/6, whose lower limit represents the threshold at which the planted bisection becomes lost amongst other random small bisections. It also remains open whether hillclimbing (i.e., annealing at temperature 0) solves the same problem; towards the latter result, Juels has recently extended our analysis and shown that random hillclimbing finds the minimum bisection with constant probability, when p - r = Omega(1) (corresponding to Delta = 2). (C) 1998 Elsevier Science B.V. All rights reserved.
引用
收藏
页码:155 / 175
页数:21
相关论文
共 27 条
[11]  
Garey M. R., 1976, Theoretical Computer Science, V1, P237, DOI 10.1016/0304-3975(76)90059-1
[12]   COOLING SCHEDULES FOR OPTIMAL ANNEALING [J].
HAJEK, B .
MATHEMATICS OF OPERATIONS RESEARCH, 1988, 13 (02) :311-329
[13]  
Jerrum M., 1988, Proceedings of the Twentieth Annual ACM Symposium on Theory of Computing, P235, DOI 10.1145/62212.62234
[14]   LARGE CLIQUES ELUDE THE METROPOLIS PROCESS [J].
JERRUM, M .
RANDOM STRUCTURES & ALGORITHMS, 1992, 3 (04) :347-359
[15]  
Jerrum M., 1993, Proceedings. 34th Annual Symposium on Foundations of Computer Science (Cat. No.93CH3368-8), P94, DOI 10.1109/SFCS.1993.366878
[16]   APPROXIMATING THE PERMANENT [J].
JERRUM, M ;
SINCLAIR, A .
SIAM JOURNAL ON COMPUTING, 1989, 18 (06) :1149-1178
[17]   OPTIMIZATION BY SIMULATED ANNEALING - AN EXPERIMENTAL EVALUATION .1. GRAPH PARTITIONING [J].
JOHNSON, DS ;
ARAGON, CR ;
MCGEOCH, LA ;
SCHEVON, C .
OPERATIONS RESEARCH, 1989, 37 (06) :865-892
[18]  
JUELS A, 1996, THESIS U CALIFORNIA
[19]   OPTIMIZATION BY SIMULATED ANNEALING [J].
KIRKPATRICK, S ;
GELATT, CD ;
VECCHI, MP .
SCIENCE, 1983, 220 (4598) :671-680
[20]  
MCDIARMID C, 1989, LOND MATH S, V141, P148