04 Research
Whether the object is a knapsack, a video stream, a sensor network, or a language model, the group's work consistently pairs rigorous algorithmic technique with systems that have to work in practice.
The group's most-cited line of work. We study the multidimensional multiple-choice knapsack problem (MMKP) — selecting one item per group under multiple resource constraints — with algorithms that scale from exact methods to distributed heuristics.
How do you guarantee quality of service when many users compete for bounded server and network capacity? This question anchored the PI's doctoral work and produced a family of admission-control and delivery techniques.
From wireless sensor networks to MPLS core networks — addressing, routing, broadcasting, and admission control under energy and bandwidth constraints.
Coordination without a coordinator: mutual exclusion, exclusion under faults, resource reservation, and architectures for ubiquitous computing.
As transistor counts grew, on-chip communication became a network problem. The group contributed simulation infrastructure and interconnection topologies for network-on-chip (NoC) design.
The group's newest thread applies machine learning and large language models to concrete problems — retrieval quality, software testing, pattern recognition, and healthcare.
→ Dig deeper
Every claim on this page traces to a published paper — browse the full annotated list, with DOI and Scholar links.
Browse publications →