1. What is the focus of research in Multi-Agent Path-Finding (MAPF)?
The focus of research in Multi-Agent Path-Finding (MAPF) includes investigating the impact of geometric constraints, expanding MAPF to include resource constraints, target assignment path-finding (TAPF), and academic problems relevant to industry. Additionally, the research extends the capabilities of standard and novel MAPF solvers to temporal graphs, explores clustering techniques, and utilizes Graph Classifications on MAPF domains to reduce computational complexity. Massively parallelised computing techniques, especially in relation to distance matrix computation and parallelised centralised MAPF, are also researched. Collaboration with Tharsus Limited allows for the application of research to novel industrial problems and the development of relevant benchmarks for Multi-Agent Systems.
read more
2. How does geometry impact MAPF problem complexity?
The underlying geometry of a MAPF problem significantly influences its complexity. Geometric invariants of a domain can lead to efficient algorithms for specific domains. In the research paper 'Atzmon et al. 2023' published at ICAPS 2023, an algorithm exploiting invariant distances between agents' start and goal locations on a MAPF problem is introduced. This algorithm's extension will include subgraphs, connectivity, and parallelized settings. The study highlights the importance of considering geometric constraints to develop scalable solutions for MAPF domains.
read more
3. What is Target Assignment Path-Finding (TAPF) and its applications?
Target Assignment Path-Finding (TAPF) is a variant of MAPF where agents have multiple target locations to visit. It has applications in warehouse automation and autonomous vehicle routing. The problem involves assigning targets to agents, ensuring each agent reaches a target location. My PhD research focuses on developing algorithms for TAPF using network flow, graph connectivity, and clustering. I aim to explore industry applications and create benchmarks for algorithm scalability.
read more
4. What are the applications of Algorithmic Graph Theory (AGT)?
Algorithmic Graph Theory (AGT) has numerous applications in various fields. It is used in network analysis, data mining, computational biology, failure and fault analysis, search, and many more. AGT provides insights into algorithms for specific graph classes like trees, planar graphs, and grid graphs. Additionally, insights from MAPF can be applied to AGT to develop novel algorithms for general graph problems, such as problems on temporal graphs, graph clustering, and the k-disjoint paths problem. Overall, AGT plays a crucial role in solving complex problems across different domains.
read more