A Unified Comparative Study of Heuristic Algorithms for Double Combinatorial Auctions: Locality-Constrained Resource Allocation Problems

摘要

Market-oriented resource allocation in cloud computing is driven by increasingly stringent needs for flexibility, fine-grained allocation, and more critically, revenue maximization. Double combinatorial auctions aptly address these demands, but their NP-hardness has hindered them from being widely adopted. Heuristic algorithms, with their input-dependent performance and solution quality, have failed to offer a robust alternative. We posit that a unifying approach for evaluating all existing algorithms, under the umbrella of a consistent problem formulation and a variety of common test cases, can propel combinatorial auctions towards real-world usage. In this paper, we performed an extensive empirical evaluation of a portfolio of heuristic algorithms for double combinatorial auctions, applied to problems with hard resource locality constraints. We found that there is no single algorithm that outperforms the others in all test scenarios. However, we offer insights into the behavior of the algorithms, and provide methods to explore the portfolio's performance over a wide range of input scenarios.