Areas of Research
Students & Supervision
Ongoing and completed supervision of Ph.D., M.Tech., and B.Sc. Honors students.
Why Distributed Algorithms?
Distributed algorithms are crucial for building scalable, robust, and efficient systems.
Scale
Add or remove components without downtime for massive scalability.
Robustness
Ensure resilience against failures, making systems fault-tolerant.
Parallelism
Leverage parallel processing to dramatically speed up computations.
Asynchrony
Allow components to work independently without tight synchronization.
Popular Problems in Distributed Computing
A taste of what research in distributed algorithms entails.
Leader Election
How can a group of robots elect a leader to coordinate tasks without a central authority?
Rendezvous Problem
Two robots must meet at a common point in an unknown environment without prior communication.
Treasure Hunt
Robots search for a treasure in a grid, learning to coordinate their movements efficiently.
Pattern Formation
Robots arrange themselves in specific geometric shapes through local interactions only.
Recommended Books
Standard references to build a solid foundation in distributed computing and mobile robotics.
Current Research Topics
Look-Compute-Move Model
The foundational framework in distributed computing for mobile robots.
- Look: Observe the environment and gather information about other robots.
- Compute: Process the information to make decisions based on local algorithms.
- Move: Act based on computations to achieve goals collaboratively.
Dispersion Problem
Scatter k mobile robots on an n-node graph ensuring each robot occupies a unique node.
Challenges
- • Robots are anonymous with limited memory
- • Synchronous or asynchronous settings
- • Directed or undirected graphs
- • Byzantine (faulty) robots can disrupt
Recent Results
Optimal time algorithms for dispersion in synchronous and asynchronous models, with solutions for directed graphs and faulty robots.
Circle Formation Problem
Arrange n autonomous mobile robots equally spaced on the circumference of a circle.
Challenges
- • Robots are anonymous and oblivious
- • No memory of previous states
- • Circle center and radius not predefined
Recent Results
Optimal algorithm in terms of colors required, with established trade-offs between time and number of colors.
Black Hole Search Problem
Locate a "black hole" (a harmful node that destroys incoming agents) in a network.
Challenges
- • Some agents can be Byzantine (malicious)
- • Agents may or may not have a network map
- • Various communication models
Recent Results
Established the number of agents needed and time complexity under various conditions, including Byzantine agents.
Black Hole Search Simulator
Try the Black Hole Search Problem hands-on: adjust parameters and watch agents search for the black hole.
Simulators & Visualizations
Interactive simulators for different mobile robot models — a good starting point for a first project.
Collaborator Network
Research is a global endeavor — co-authors from institutions around the world.
Ready to Start Your Research Journey?
If you're passionate about distributed algorithms and want to explore research with me, please include the essentials that make a first discussion productive.
CV or academic resume
Relevant coursework and grades
Research interests or problem areas
Prior projects, papers, or code links