Bayesian Optimization-Based Combinatorial Assignment

Authors: Jakob Weissteiner, Jakob Heiss, Julien Siems, Sven Seuken

AAAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We run computational experiments in several spectrum auction domains to evaluate BOCA s performance. Our results show that BOCA achieves higher allocative efficiency than state-of-the-art approaches.
Researcher Affiliation Academia 1University of Zurich 2ETH Zurich 3ETH AI Center
Pseudocode No The information is insufficient. The paper describes various algorithms and methods but does not include any explicitly labeled "Pseudocode" or "Algorithm" blocks, nor does it present structured code-like steps for a procedure.
Open Source Code Yes Our source code is publicly available on Git Hub via: https://github.com/marketdesignresearch/BOCA.
Open Datasets Yes To generate synthetic CA instances, we use the following three domains from the spectrum auction test suite (SATS) (Weiss, Lubin, and Seuken 2017): LSVM, SRVM, and MRVM (see Appendix G.1 for details).
Dataset Splits No The information is insufficient. The paper mentions training on 'Dtrain' and evaluating on a 'disjoint test set Dtest' but does not provide specific split percentages or counts for training, validation, and test sets to reproduce the data partitioning.
Hardware Specification No The information is insufficient. The paper does not provide specific hardware details such as GPU/CPU models, processor types, or memory amounts used for running its experiments.
Software Dependencies No The information is insufficient. While the paper mentions using a MILP solver (e.g., implied for solving Equation 10 and Theorem 2), it does not specify any software dependencies with version numbers (e.g., 'CPLEX 12.4' or 'PyTorch 1.9').
Experiment Setup Yes To enable a fair comparison against prior work, for each domain, we use Qinit = 40 initial random queries (including the full bundle for the calculation of M100%-u UB i ) and set the query budget to Qmax = 100... We use random search (RS) (Bergstra and Bengio 2012) to optimize the hyperparameters of the mean MVNN Mmean i and of our MVNN-based u UB Mu UB i . The HPO includes the NNarchitecture parameters, training parameters, NOMU parameters, and initialization parameters (see Section 3.2).