Online Convex Optimization with Stochastic Constraints

Authors: Hao Yu, Michael Neely, Xiaohan Wei

NeurIPS 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on a real-world data center scheduling problem further verify the performance of the new algorithm. (Abstract) ... Fig. 1(c)(d) plot the performance of 4 algorithms, where the running average is the time average up to the current slot. Fig. 1(c) compares electricity cost while Fig. 1(d) compares unserved jobs.
Researcher Affiliation Academia Hao Yu, Michael J. Neely, Xiaohan Wei Department of Electrical Engineering, University of Southern California {yuhao,mjneely,xiaohanw}@usc.edu
Pseudocode Yes Algorithm 1 Let V > 0 and > 0 be constant algorithm parameters. Choose x(1) 2 X0 arbitrarily and let Qk(1) = 0, 8k 2 {1, 2, . . . , m}. At the end of each round t 2 {1, 2, . . .}, observe f t( ) and gt( ) and do the following: Choose x(t + 1) that solves ... Update each virtual queue Qk(t + 1), 8k 2 {1, 2, . . . , m} via...
Open Source Code No The paper does not provide an explicit statement or link indicating that the source code for the described methodology is publicly available.
Open Datasets Yes We use ci(t) from real-world 5-minute average electricity price data at 10 different zones in New York city between 05/01/2017 and 05/10/2017 obtained from NYISO [1]. [1] New York ISO open access pricing data. http://www.nyiso.com/.
Dataset Splits No The paper does not specify exact training, validation, or test dataset splits. It describes using real-world data and plotting 'running average' performance over time.
Hardware Specification No The paper describes the data center infrastructure for the experiment (100 geographically distributed servers, 10 clusters) but does not specify the hardware used to run the simulations or experiments (e.g., CPU/GPU models, memory, etc.).
Software Dependencies No The paper does not provide specific software dependencies or their version numbers for replicating the experiments.
Experiment Setup No The paper describes the job scheduling problem and the data used, but it does not provide specific hyperparameters for Algorithm 1 (e.g., learning rates, batch sizes, number of epochs) or detailed system-level training settings beyond the time slot (5 minutes).