TSPDiffuser: Diffusion Models as Learned Samplers for Traveling Salesperson Path Planning Problems
This work addresses path planning for robotics or logistics in complex environments, representing an incremental advance by applying diffusion models to a known bottleneck in TSPPPs.
The paper tackles the traveling salesperson path planning problem (TSPPP) in obstacle-rich environments by introducing TSPDiffuser, a diffusion model trained on solution data to generate plausible paths and estimate travel costs efficiently, achieving improved trade-offs between solution quality and computational time in synthetic and real-world evaluations.
This paper presents TSPDiffuser, a novel data-driven path planner for traveling salesperson path planning problems (TSPPPs) in environments rich with obstacles. Given a set of destinations within obstacle maps, our objective is to efficiently find the shortest possible collision-free path that visits all the destinations. In TSPDiffuser, we train a diffusion model on a large collection of TSPPP instances and their respective solutions to generate plausible paths for unseen problem instances. The model can then be employed as a learned sampler to construct a roadmap that contains potential solutions with a small number of nodes and edges. This approach enables efficient and accurate estimation of travel costs between destinations, effectively addressing the primary computational challenge in solving TSPPPs. Experimental evaluations with diverse synthetic and real-world indoor/outdoor environments demonstrate the effectiveness of TSPDiffuser over existing methods in terms of the trade-off between solution quality and computational time requirements.