I am thrilled to announce that Avi Gal and I have received a competitive grant from the Israel Science Foundation (ISF) for 2024-2028. This funding supports our work on AI-driven algorithms to detect data deviations and improve prediction models and sensors. We are grateful to the ISF for their support and confidence in our research. We recruit new graduate students.
Scheduling for Cloud Computing
We analyze the performance of a simple scheduling rule for cloud computing settings in which different jobs can run in parallel on the same server (machine). The suggested algorithms can easily be used for offline (static) and online scheduling with very good performance. This paper was just accepted to Annals of Operations Research (AOR) – look at it here. See my other recent publications.
Process Detection from Video
Winning a DFG Grant
I thank the DFG for supporting my research via a competitive grant for the years 2023-2026.
The German Research Foundation (DFG) supports mine and my partners’ research: Shimrit Shtern (Technion, Haifa), Erwin Pesch (Siegen University, Siegen), and Dominik Kress (Helmut Schmidt University/University of the Federal Armed Forces Hamburg).
Machine Learning with Resource Constraints
Want to use ML to allocate a limited amount of resources (e.g., vaccines) between multiple entities (e.g., patients)? Our paper, accepted to Engineering Applications of Artificial Intelligence (IF=7.8), suggests a machine learning approach that inherently incorporates the resource constraints into the learning process. Read it in this link, especially for you.
Written with our talented MSc student Danit Abukasis Shifman from Bar-Ilan Faculty of Engineering IISE Track, Kejun Huang, Xiaochen Xian, and Gonen Singer.
Queueing Theory vs. Machine Learning
Our paper “Delay Prediction for Managing Multi-Class Service Systems: An Investigation of Queueing Theory and Machine Learning Approaches” was accepted to IEEE Transactions on Engineering Management (impact factor: 8.7).
Read it to know when to prefer the simple queueing theory formulas over machine learning models, and when the latter models are worthwhile.
Contact me for more details.
Professor Paul Feigin and the talented Elisheva Chocron (MSc.) are my co-authors.
Read the abstract and continue to the full paper:
Customer waiting time prediction is key to managing service systems. Predicting how long a customer will wait for service at the time of their arrival can provide important information to the customer and serve as a tool for the operations manager.
Recent studies that suggested machine learning algorithms for waiting time prediction as an alternative to the standard queueing theory approaches investigated specific systems with mixed results regarding the superiority of a particular approach.
We provide a systematic investigation of common violations of queueing theory assumptions on waiting time prediction in the context of single-queue many-server systems. These violations include non-stationarity, non-exponential service times, state-dependent service times, abandonments, and customers with different priorities. Using different machine learning models as well as queueing theory based methods, we seek to determine under what regimes machine learning prediction is to be preferred to queueing theory based predictors.
Our results suggest that queueing theory models produce comparable and frequently better predictions versus machine learning algorithms at a much lower computational cost. Under other assumptions, such as high priority for a specific type of customer, machine learning predictions may outperform queueing theory predictions. Our results may guide the selection of a delay prediction approach for service systems. Read more…
Winning a Personal ISF Grant
Between Capacitated Machines, Ovens and Cloud Computing (accepted to WAOA21)
Interestingly, problems with capacitated machines were initially encountered in production settings where jobs are processed in batches such as scheduling jobs for heat treatment ovens and wafer fabrication processes. The interest in these problems has increased because of its relevance for modeling modern cloud computing environments. Indeed, differently than most scheduling models in which a resource serves a single job at any given time, in modern cloud computing environments, multiple jobs can run concurrently on the same server subject to its capacity constraints (e.g., memory, cores, bandwidth). Read our paper with new results about the problem. For more, refer to my publications page. This paper was written with my talented MSc. student Iyar Zaks and colleague Dr. Ilan Cohen.
תואר אוניברסיטאי חדש בהנדסת תעשייה ומערכות מידע
my latest paper about scheduling via adaptive robust optimization
Please Share, Cite and Comment. With Krzysztof Postek and Shimrit Shtern. link Video Presentation