English

Balanced Allocation on Graphs: A Random Walk Approach

Data Structures and Algorithms 2016-03-01 v4 Discrete Mathematics Probability

Abstract

In this paper we propose algorithms for allocating nn sequential balls into nn bins that are interconnected as a dd-regular nn-vertex graph GG, where d3d\ge3 can be any integer.Let ll be a given positive integer. In each round tt, 1tn1\le t\le n, ball tt picks a node of GG uniformly at random and performs a non-backtracking random walk of length ll from the chosen node.Then it allocates itself on one of the visited nodes with minimum load (ties are broken uniformly at random). Suppose that GG has a sufficiently large girth and d=ω(logn)d=\omega(\log n). Then we establish an upper bound for the maximum number of balls at any bin after allocating nn balls by the algorithm, called {\it maximum load}, in terms of ll with high probability. We also show that the upper bound is at most an O(loglogn)O(\log\log n) factor above the lower bound that is proved for the algorithm. In particular, we show that if we set l=(logn)1+ϵ2l=\lfloor(\log n)^{\frac{1+\epsilon}{2}}\rfloor, for every constant ϵ(0,1)\epsilon\in (0, 1), and GG has girth at least ω(l)\omega(l), then the maximum load attained by the algorithm is bounded by O(1/ϵ)O(1/\epsilon) with high probability.Finally, we slightly modify the algorithm to have similar results for balanced allocation on dd-regular graph with d[3,O(logn)]d\in[3, O(\log n)] and sufficiently large girth.

Keywords

Cite

@article{arxiv.1407.2575,
  title  = {Balanced Allocation on Graphs: A Random Walk Approach},
  author = {Ali Pourmiri},
  journal= {arXiv preprint arXiv:1407.2575},
  year   = {2016}
}
R2 v1 2026-06-22T04:59:52.396Z