##
###
LCL Problems on Grids

Sebastian Brandt, Juho Hirvonen, Janne H. Korhonen, Tuomo Lempiäinen, Patric R.J. Östergård, Christopher Purcell, Joel Rybicki, Jukka Suomela, Przemysław Uznański

2017
*
Proceedings of the ACM Symposium on Principles of Distributed Computing - PODC '17
*

LCLs or locally checkable labelling problems (e.g. maximal independent set, maximal matching, and vertex colouring) in the LOCAL model of computation are very well-understood in cycles (toroidal 1-dimensional grids): every problem has a complexity of O(1), Θ(log * n), or Θ(n), and the design of optimal algorithms can be fully automated. This work develops the complexity theory of LCL problems for toroidal 2-dimensional grids. The complexity classes are the same as in the 1-dimensional case:
## more »

... , Θ(log * n), and Θ(n). However, given an LCL problem it is undecidable whether its complexity is Θ(log * n) or Θ(n) in 2-dimensional grids. Nevertheless, if we correctly guess that the complexity of a problem is Θ(log * n), we can completely automate the design of optimal algorithms. For any problem we can find an algorithm that is of a normal form A • S k , where A is a finite function, S k is an algorithm for finding a maximal independent set in kth power of the grid, and k is a constant. Finally, partially with the help of automated design tools, we classify the complexity of several concrete LCL problems related to colourings and orientations. arXiv:1702.05456v2 [cs.DC] 24 May 2017 1.1 Problem setting: LCL problems on grids 92 33 77 57 49 26 71 79 8 62 48 24 31 21 15 30 60 67 0 5 17 95 23 47 87 80 25 38 20 64 45 61 91 51 69 1 74 55 3 98 88 99 58 53 63 40 16 2 39 Grids. In this work, we study distributed algorithms in a setting where the underlying input graph is a grid. Specifically, we consider the complexity of locally checkable labelling problems, or LCL problems, in the standard LOCAL model of distributed complexity, and consider graphs that are toroidal two-dimensional n × n grids with a consistent orientation; we focus on the two-dimensional case for concreteness, but most of our results generalise to d-dimensional grids of arbitrary dimensions. This setting occupies a middle ground between the wellunderstood directed n-cycles [10, 32] , where all solvable LCL problems are known to have deterministic time complexity either O(1), Θ(log * n) or Θ(n), and the more complicated setting of general n-vertex graphs, where intermediate problems with time complexities such as Θ(log n) are known to exist, even for bounded-degree graphs. Grid-like systems with local dynamics also occur frequently in the study of real-world phenomena. However, grids have so far not been systematically studied from a distributed computing perspective. LOCAL model and LCL problems. In the LOCAL model of distributed computing, nodes are labelled with unique numerical identifiers with O(log n) bits. A time-t algorithm in this model is simply a mapping from radius-t neighbourhoods to local outputs; equivalently, it can be interpreted as a message-passing algorithm in which the nodes exchange messages for t synchronous rounds and then announce their local outputs. LCL problems are graph problems for which the feasibility of a solution can be verified by checking the solution for each O(1)-radius neighbourhood; if all local neighbourhoods look valid, the solution is also globally valid. Examples of such problems include vertex colouring, edge colouring, maximal independent sets, and maximal matchings. We refer to Section 3 for precise definitions. Example: colouring the grid. To illustrate the type of questions we are interested in this work, consider k-colouring on n × n grids. For k = 2, the problem is inherently global with complexity Θ(n), while colouring any graph of maximum degree ∆ = 4 with ∆ + 1 = 5 colours can be done in O(log * n) rounds. But what about k = 3 and k = 4? In particular, does either of these have an intermediate (polylogarithmic) complexity, as is known to happen with ∆-colouring on general bounded-degree graphs [10, 34]? We will see that neither 3-colouring nor 4-colouring is intermediate on grids: 3-colouring requires Θ(n) rounds, while 4-colouring can be solved in O(log * n) rounds.

doi:10.1145/3087801.3087833
dblp:conf/podc/BrandtHKLOPRSU17
fatcat:q6vegq37bbgvpji7tuegyhbp6e