In the last decade, the complexity of networks has increased significantly to accommodate the rise of innovative applications. This growing complexity has rendered traditional human-in-the-loop network management approaches inadequate, necessitating ...

In the realm of stochastic control, particularly in the fields of economics and engineering, Markov Decision Processes (MDP's) are employed to represent various processes ranging from asset management to transportation logistics. Upon closer examination ...

Data center workloads are composed of multiresource jobs requiring a variety of computational resources including CPU cores, memory, disk space, and hardware accelerators. Modern servers can run multiple jobs in parallel, but a set of jobs can only run ...

A well-designed scheduling policy can significantly improve the performance of a queueing system, without requiring any additional resources. While scheduling is well-understood in the single-server setting, much less is known in the multiserver setting. ...

Preemptive scheduling policies, which allow pausing jobs mid-service, are ubiquitous because they allow important jobs to receive service ahead of unimportant jobs that would otherwise delay their completion. The canonical example is Shortest Remaining ...

We study the design of risk-sensitive online algorithms, in which risk measures are used in the competitive analysis of randomized online algorithms.

In this paper, we initiate the study of an online conversion problem that incorporates group fairness guarantees. This problem aims to distribute a resource with fixed capacity to a sequence of buyers based on their offered prices. Each buyer belongs to ...

Multidimensional Markov processes arise in many aspects of the mathematical performance analysis, modeling and optimization of computer systems and networks. Within this context, general classes of structured Markov processes are of particular importance ...

In this study, we investigate the performance of multi-armed bandit algorithms in environments characterized by heavytailed and non-stationary reward distributions, a setting that deviates from the conventional risk-neutral and sub- Gaussian assumptions.

Recently it was shown that the response time of First-Come- First-Served (FCFS) scheduling can be stochastically and asymptotically improved upon by the Nudge scheduling algorithm in case of light-tailed job size distributions. Such improvements are ...

Motivated by Q-learning, we study nonsmooth contractive stochastic approximation (SA) with constant stepsize. We focus on two important classes of dynamics: 1) nonsmooth contractive SA with additive noise, and 2) synchronous and asynchronous Q-learning, ...

We consider large-scale load balancing systems where processing time distribution of tasks depend on both task and server types. We analyze the system in the asymptotic regime where the number of task and server types tend to infinity proportionally to ...

This paper studies the Heavy Traffic (HT) joint distribution of queue lengths in an Input-queued switch (IQ switch) operating under the MaxWeight scheduling policy. IQ switchserve as representative of SPNs that do not satisfy the socalled Complete ...

We consider the classic problem of preemptively scheduling jobs in a queue to minimize mean number-in-system, or equivalently mean response time. Even in single-server queueing models, this can be a nontrivial problem whose answer depends on the ...

Brief Biography: Md Kamran Chowdhury Shisher is a final-year Ph.D. candidate in the Department of Electrical and Computer Engineering at Auburn University, Auburn, AL, USA. He expects to graduate by May 2024. His Ph.D. works are advised by Yin Sun. ...

Brief Biography: Abhin Shah is a final-year Ph.D. student in the department of Electrical Engineering and Computer Science atMIT, where he is a recipient ofMIT's Jacobs Presidential Fellowship. He has interned with Google Research (2021) and IBM Research ...

Brief Biography: Tasmeen Zaman Ornee is a PhD candidate at the Electrical and Computer Engineering Department of Auburn University. She joined the Real-time Networking lab in Fall 2018. Her PhD supervisor is Prof. Yin Sun. She completed her M.Sc. in ...

Sequential decision-making algorithms are ubiquitous in the design and optimization of large-scale systems due to their practical impact. The typical algorithmic paradigm ignores the sequential notion of these problems: use a historical dataset to ...

Stochastic Processing Networks (SPNs) [24] are ubiquitous in engineering with applications in Data Centers (eg. packet routing), Telecommunication (eg. call centers), Transportation, etc. Stochastic Matching Networks (SMNs) are newly emerging variation ...

Scheduling theory is a key tool for reducing latency (i.e. response time) in queueing systems. Scheduling, i.e. choosing the order in which to serve jobs, can reduce response time by an order of magnitude with no additional resources. Scheduling theory ...