Projects

1. Approximation Algorithms for Bipartite Maximum Edge Weight Matching: The goal of this research is to develop faster (parallel) algorithms for bipartite edge weighted matching problem. Since exact algorithms for this problem are computationally expensive and very hard parallelize, we look into different heuristics to develop approximation algorithms which are both reasonable in quality and easier to parallelize.

2. Spectral Unmixing for Processing the Flow Cytometric Data: The research objective is to analyze flow cytometric data of white blood cells to separate cancerous (Leukemic) blood cells from the healthy cells. We employ techniques from remote sensing along with constrained non linear optimization algorithms to de-convolute spectrally overlapping flow cytometric data.

3. Multithreaded Maximum Cardinality Matching: The research objective is to design parallel multithreaded algorithm for maximum cardinality matching of bipartite graphs. We design the multithreaded algorithms and do the performance analysis against the best available serial counterpart on different machine architectures.

4. Spatial Reasoning: Many applications benefit from the inference of new knowledge about the spatial (e.g., Topological, Cardinal Directions) relationships between spatial objects on the basis of already available and explicit spatial information. Hence, our goal is to build a reasoning model by means of which we can infer spatial knowledge from the existing spatial facts.

5. AIST: The research objective is the development of Moving Objects Database technology to efficiently store and retrieve both non spatial and moving object data. Currently the main focus is on the analysis and faster access of hurricane data for better forecasting.

6. Community Detection through Encounter Profiling: The project was on determining the underlying community of human mobility behavior through the encounters they make to each other independently throughout their daily schedule and eventually take advantage of that information for routing in Delay Tolerant Networks.