Balanced Allocation on Graphs: A Random Walk Approach

TL;DR

This paper introduces a random walk-based algorithm for balanced load allocation on regular graphs, achieving near-optimal maximum load bounds under certain girth and degree conditions.

Contribution

It presents a novel random walk approach for load balancing on graphs with theoretical bounds, extending previous models to broader graph classes.

Findings

Maximum load is bounded by O(1/ε) with high probability.

Algorithm performs well on graphs with large girth and degree d=ω(log n).

Modified algorithm achieves similar bounds for degrees up to O(log n).

Abstract

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