Resource-sharing system scheduling and circular chromatic number

被引:10
|
作者
Yeh, HG [1 ]
Zhu, XD
机构
[1] Natl Cent Univ, Dept Math, Taoyuan 320, Taiwan
[2] Natl Sun Yat Sen Univ, Dept Appl Math, Kaohsiung 804, Taiwan
关键词
circular chromatic number; scheduling; fairness; fractional chromatic number; homomorphism; scheduling by edge reversal; minimum mean cycle;
D O I
10.1016/j.tcs.2004.12.005
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
A graph G is used as a model for a resource sharing system, where each vertex represents a process and an edge joining two vertices means that the corresponding processes share a resource. A scheduling of G is a mapping f : {1, 2, 3,...} -> 2(V(G)), where f (i) consists of processes that are operating at round i. The rate of f is defined as rate(f) = lim sup(n ->infinity) Sigma(i=1)(n) vertical bar f(i)vertical bar/n vertical bar V (G)vertical bar, which is the average fraction of operating processes at each round. A scheduling is fair if adjacent vertices alternate their turns in operating. The operating rate gamma* (G) of G is the maximum rate of a fair scheduling. Fair schedulings of a graph was first studied by Barbosa and Gafni. They introduced the method of "scheduling by edge reversal" which derives a fair scheduling through an acyclic orientation. Through scheduling by edge reversal, Barbosa and Gafni related gamma* (G) to the structure of acyclic orientations of G. We point out that this relation implies that gamma*(G) is equal to the reciprocal of the circular chromatic number of G. Both circular coloring and scheduling by edge reversal have been studied extensively in the past decade. The former by graph theorists, and the latter by computer scientists. However, it seems that neither side knew the existence of the other. This paper intends to build a connection between the two sides. We show that certain open problems concerning scheduling by edge reversal are indeed solved under the language of circular coloring. In the study of fair scheduling, Barbosa and Gafni defined a variation of multiple coloring of graphs: the interleaved p-color, q-tuple colorings. We formulate the interleaved coloring as a graph homomorphism problem. In the study of circular chromatic number, Bondy and Hell defined (p, q)-colorings and also formulated it as a graph homomorphism problem. We prove that the target graph for the interleaved p-color, q-tuple coloring and the target graph of (p, q)-coloring are homomorphically equivalent. This gives another proof of the fact that gamma*(G) = 1/chi(C)(G). Moreover, the proof gives an explicit formula which deduces an optimal circular coloring of G from an optimal interleaved coloring of G, and vice versa. This paper also introduces two other schedulings of a graph, the weakly fair scheduling and the strongly fair scheduling. It is proved that the rate of an optimal strongly fair scheduling of a graph G is also equal to the reciprocal of the circular chromatic number of G, and the rate of an optimal weakly fair scheduling of G is equal to the reciprocal of the fractional chromatic number of G. Barbosa and Gafni presented an algorithm that determines the rate gamma(o) (w) of the scheduling induced by an acyclic orientation w of G. By using Karp's minimum mean cycle algorithm, we give a faster algorithm to calculate gamma(o)(w). (c) 2005 Elsevier B.V. All rights reserved.
引用
收藏
页码:447 / 460
页数:14
相关论文
共 50 条
  • [41] The circular chromatic number of hypergraphs
    Brewster, Richard C.
    MacGillivray, Gary
    Shepherd, Laura
    [J]. DISCRETE MATHEMATICS, 2009, 309 (18) : 5757 - 5765
  • [42] The circular chromatic number of a digraph
    Bokal, D
    Fijavz, G
    Juvan, M
    Kayll, PM
    Mohar, B
    [J]. JOURNAL OF GRAPH THEORY, 2004, 46 (03) : 227 - 240
  • [43] A new trust framework for resource-sharing in the grid environment
    Hu, HL
    Chen, D
    Huang, CQ
    [J]. COMPUTATIONAL SCIENCE - ICCS 2005, PT 3, 2005, 3516 : 221 - 228
  • [44] CARMA: Collective Adaptive Resource-sharing Markovian Agents
    Bortolussi, Luca
    De Nicola, Rocco
    Galpin, Vashti
    Gilmore, Stephen
    Hillston, Jane
    Latella, Diego
    Loreti, Michele
    Massink, Mieke
    [J]. ELECTRONIC PROCEEDINGS IN THEORETICAL COMPUTER SCIENCE, 2015, (194): : 16 - 31
  • [45] Resource-sharing and service deployment in virtual data centers
    Graupner, S
    Kotov, V
    Trinks, H
    [J]. 22ND INTERNATIONAL CONFERENCE ON DISTRIBUTED COMPUTING SYSTEMS WORKSHOP, PROCEEDINGS, 2002, : 666 - 671
  • [46] Resource-sharing in multiple-component working memory
    Doherty, Jason M.
    Logie, Robert H.
    [J]. MEMORY & COGNITION, 2016, 44 (08) : 1157 - 1167
  • [47] Optimal control of a resource-sharing multiprocessor with periodic maintenance
    Kogan, K
    Lou, S
    Herbon, A
    [J]. IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2002, 47 (08) : 1342 - 1346
  • [48] Circular business model implementation: Design choices, orchestration strategies, and transition pathways for resource-sharing solutions
    Palmie, Maximilian
    Boehm, Jonas
    Lekkas, Charlotte-Katharina
    Parida, Vinit
    Wincent, Joakim
    Gassmann, Oliver
    [J]. JOURNAL OF CLEANER PRODUCTION, 2021, 280
  • [49] APPLICATION OF MODERN TECHNOLOGIES TO INTERLIBRARY RESOURCE-SHARING NETWORKS
    REINTJES, JF
    [J]. JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE, 1984, 35 (01): : 45 - 52
  • [50] Resource-sharing roles and responsibilities for CISTI: change is the constant
    Krym, N
    VanBuskirk, M
    [J]. INTERLENDING & DOCUMENT SUPPLY, 2001, 29 (01): : 11 - 16