Operations Research 2025
Abstract Submission

72. Heuristics for Variable Cost and Size Cluster Vector Bin Packing (VCSCVBP)

Invited abstract in session FA-2: Scheduling & Packing, stream Discrete and Combinatorial Optimization.

Friday, 8:45-10:15
Room: H4

Authors (first author is the speaker)

1. Laura Wolf
Institute for Operations Research (IOR), Karlsruhe Institute of Technology (KIT)
2. Sabrina Klos
Technische Hochschule Würzburg-Schweinfurt
3. Stefan Nickel
Institute for Operations Research (IOR), Karlsruhe Institute of Technology (KIT)

Abstract

Vector bin packing, a multidimensional extension of the bin packing problem, involves packing a set of d-dimensional vectors, each representing the resource requirements of an item, into a d-dimensional vector representing the available bin resources. Motivated by the increasing demand for computing resources and the need for efficient capacity planning for cloud infrastructure, we present a model for vector bin packing with bins of variable sizes organized into clusters, where costs are incurred for utilizing clusters (VCSCVBP). The novelty here lies in the cluster structure of bins. We propose and evaluate several heuristics to address this problem, considering both homogeneous cluster structures (a single bin type per cluster) as well as heterogeneous structures (multiple bin types per cluster). Computational experiments demonstrate significant runtime improvements with only minor deviations from the best solutions found by the solver.

Keywords

Status: accepted

Back to the list of papers