Selected Videos

Strongly Tail-Optimal Scheduling in the Light-Tailed M/G/1


Recent Progress in Queueing and Scheduling Theory (for a TCS Audience)

Talks

New Scheduling and Dispatching Algorithms for Optimizing Tail Latency

Stochastic Networks 2026, Forty Years of Queueing Systems 2026, CMU Tepper OM Seminar 2026, Google Algorithms Seminar 2025, Toronto Rotman Young Scholar Seminar 2025.
[slides]
Previous title: New Tools for Optimizing Tail Latency in Queues: Generating Functions, the Gittins Index, and Gurobi

What TCS Can Do for Queueing and What Queueing Can Do for TCS in Scheduling Theory

Schloss Dagstuhl Online Algorithms Beyond Competitive Analysis 2025.
[slides]

The Gittins Index: A Design Principle for Decision Making Under Uncertainty

Tutorial at INFORMS 2025 with Alex Terenin.
Parts 1 & 3: Pandora’s box & tail scheduling [slides]
Part 2: Bayesian optimization [slides]
[paper]

Scheduling and dispatching problems in queues: recently closed, hopelessly open, and promisingly ajar

INFORMS APS 2025.
[slides]

A New Design Tool for Bayesian Optimization: The Gittins Index

INFORMS 2025, SC&RL 2025, Cambridge CBL Seminar 2024.
[slides] [paper]

What Problems Does the Gittins Index Solve?

BIRS Dynamic Allocation and Matching 2025.
[slides] [video]

Local Hedging Solves Pandora’s Box Problems with Nonobligatory Inspection

INFORMS 2024, UC Berkeley Simons Institute 2024, Cornell CS Theory Seminar 2023, MIT Algorithms & Complexity Seminar 2023.
[slides] [paper]
Previous title: Local Hedging Solves Pandora’s Box Problems with Optional Inspection.

Strongly Tail-Optimal Scheduling in the Light-Tailed M/G/1

YinzOR 2024, SPCOM 2024, SNAPP Seminar 2024, UMass Amherst CS Theory Seminar 2024.
[slides] [video] [paper]

Performance of the Gittins Policy in the G/G/1 and G/G/k, With and Without Setup Times

INFORMS APS 2023.
[slides] [paper]

The Role of Advanced Math in Teaching Performance Modeling

TeaPACS 2023.
[slides] [paper]

Recent Progress in Queueing and Scheduling Theory (for a TCS Audience)

TCS+ 2023.
[slides] [video]

How Robust Is the Gittins Policy for Queue Scheduling?

JMM 2023, UC Berkeley Simons Institute 2022.
[slides] [video]
Previous title: How Robust Is the Gittins Index? A Queueing Case Study

Markov-Process Multi-Golf: A Gittins Index Tutorial

UC Berkeley Simons Institute Whiteboard Talk 2022.
[notes]

A New Toolbox for Scheduling Theory

Stanford GSB OIT Seminar 2023, Harvard Systems-Theory Seminar 2023, Northwestern Kellog Seminar 2022, Amsterdam Applied Probability Seminar 2022, TU/e SOR Seminar 2022, Michigan IOE Seminar 2022, Cornell ORIE Special Seminar 2022, Chicago Booth Workshop in Operations/Management Science 2022, Caltech Frontiers in Computing + Mathematical Sciences Symposium 2022.
[slides] [thesis]

Drifting Towards Progress in Multiserver Scheduling

CORS/INFORMS International 2022.
[slides]

WINE: A New Queueing Identity for Analyzing Scheduling Policies in Multiserver Systems

Cornell CAM Seminar 2023, CORS/INFORMS International 2022, INFORMS 2021.
[slides]
Previous title: A New Queueing Identity with Applications to Multiserver Scheduling.

Uniform Bounds for Scheduling with Job Size Estimates

LATA 2023, ITCS 2022.
[slides] [video] [paper]

The Gittins Policy in the M/G/1 Queue

EURO 2022, WiOpt 2021.
[slides] [video] [paper]

How to Schedule Near-Optimally under Real-World Constraints

Michigan CSP Seminar 2021.
[slides] [paper]

Coupling Techniques for Complex Control Problems

Tutorial at SIGMETRICS 2021 with Sid Banerjee.
Part 1: sample-path coupling [slides] [video]
Part 2: steady-state coupling [slides] [video]

The Gittins Policy Is Nearly Optimal in the M/G/k under Extremely General Conditions

SIGMETRICS 2021, Stanford Information Theory Forum 2021, UIUC SINE Seminar 2021.
[slides] [video] [paper] [long slides]

Bounding Mean Slowdown in Multiserver Systems

MAMA 2021.
[slides] [paper]

Near-Optimal Scheduling: Towards a Unified Theory

Ohio State Seminar 2021.
[slides]

Optimal Multiserver Scheduling with Unknown Job Sizes in Heavy Traffic

MAMA 2020.
[slides] [video] [paper]

Characterizing Policies with Optimal Response Time Tails under Heavy-Tailed Job Sizes

SIGMETRICS 2020.
[slides] [video] [paper]

Simple Near-Optimal Scheduling for the M/G/1

SIGMETRICS 2020, MAMA 2019, INFORMS 2019, Stanford ISL Colloquium 2019.
SIGMETRICS 2020 Best Video Award.
[slides] [video] [paper]

Unfair Scheduling Patterns in NUMA Architectures

PACT 2019.
[slides] [paper]

The Power of SOAP Scheduling

Tutorial at SIGMETRICS 2019 with Mor Harchol-Balter.
[slides]

Optimal Scheduling and Exact Response Time Analysis for Multistage Jobs

INFORMS 2018, YEQT 2018.
[slides] [paper]

SOAP: One Clean Analysis of All Age-Based Scheduling Policies

INFORMS 2018, SIGMETRICS 2018, CMU CS Theory Lunch 2018, TU/e Stochastics Colloquium 2018, UvA SPIP Meeting 2018, IBM Research Seminar 2019, UW CS Theory Seminar 2019, UC Berkeley IEOR Seminar 2019.
[slides] [video] [paper] [short slides]

Open Problem—M/G/1 Scheduling with Preemption Delays

INFORMS 2018.
[slides] [paper]

Optimally Scheduling Jobs with Multiple Tasks

INFORMS 2017, YinzOR 2017, INFORMS APS 2017, MAMA 2017.
[slides] [paper]

A Program Optimization for Automatic Database Result Caching

POPL 2017, PLDI 2016 Student Research Competition.
[slides] [paper]

Scheduling with the Gittins Index

CMU CS Theory Lunch 2017.
[slides] [video]