Enhancing Constraint-Based Multi-Objective Combinatorial Optimization

Authors: Miguel Terra-Neves, Inês Lynce, Vasco Manquinho

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, experimental results on the Virtual Machine Consolidation (VMC) problem show the effectiveness of the proposed techniques. In this section, the performance of the techniques proposed in sections 3 and 4 is evaluated on instances of the Virtual Machine Consolidation (VMC) problem.
Researcher Affiliation Academia Miguel Terra-Neves, Inês Lynce, Vasco Manquinho INESC-ID / Instituto Superior Técnico, Universidade de Lisboa, Portugal
Pseudocode Yes Algorithm 1: CLD algorithm for computing an MCS; Algorithm 2: Sample MCS algorithm for sampling a single MCS of a MOCO instance; Algorithm 3: MCSEnum PD algorithm for enumerating MCSs of a MOCO instance using path diversification
Open Source Code Yes The benchmark set and the prototype that implements the algorithms evaluated in this paper are publicly available online2. (2http://sat.inesc-id.pt/dome)
Open Datasets Yes The evaluation is performed on the VMC benchmarks used in the work of Terra-Neves, Lynce, and Man quinho(2017), which are based on subsets of workload traces randomly selected from the Google Cluster Data project1. (1http://code.google.com/p/googleclusterdata/)
Dataset Splits No The paper describes the benchmark instances and their characteristics, but it does not specify explicit train/validation/test dataset splits, percentages, or methodology for partitioning the data into these sets.
Hardware Specification Yes The evaluation was conducted on an AMD Opteron 6376 (2.3 GHz) with 128 GB of RAM.
Software Dependencies Yes All algorithms were implemented in Java and Sat4j-PB (Le Berre and Parrain 2010) (version 2.3.4) was used as the PBS solver.
Experiment Setup Yes Each algorithm was executed with a memory limit of 4 GB and a time limit of 1800 seconds. Randomized algorithms were executed with 10 different seeds for each instance, and the analysis is performed using the median values over all executions. The m (n P ) parameter of Sample MCS (MCSEnum PD) was set to 1 (4), as suggested by our empirical evaluation. MGGA was adapted to consider migration costs instead and was configured to use a population size of 12, and crossover rate and mutation rate as suggested by Xu and Fortes(2010).