Multi-Agent Pick and Delivery with Capacities: Action Planning vs Path Finding
summary
Defines MAPDC: robots pick and deliver items with per-robot capacity limits, encoded two ways in Answer Set Programming — a time-annotated action/planning encoding and a traversal/path-finding encoding — run with single-shot and multi-shot Clingo. Benchmarks on warehouse grids with shelves show capacity increases (e.g., 1→2 items) cut makespan and CPU time more than adding agents, tasks 16→32 cause timeouts, and multi-shot helps the planning encoding (500s→~300s) while sometimes degrading traversal performance.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 0:48Motivation
- 2:17Order Picking
- 2:55Map Problem
- 5:22Approach
- 6:20PlanBased Solution
- 6:49Action Domain
- 9:56Output
- 11:56ASP Encoding
- 14:52Picking and Delivery
- 18:15Input
- 19:09ASp
- 19:52Experimental Setup
- 23:21Capacity Results
- 24:51Conclusion
related talks
40mCo-operative robots sharing the load
26mPAARL: An Interactive System for Policy-Aware Planning in Autonomous Agents(…)
57mFrom Shop Floor Control to Physical AI: Bridging Industrial Connectivity and Intelligent Automation
29mEfficient Path Planning for Search and Rescue Missions Using Neural Networks
01h 24mBuilding Multiagent Systems: An Ontology-Driven Approach with SPADE
18mRobotics: why now?
21mDistributed Reinforcement Learning for Robotic Assembly
22mAutonomous Delivery and Multi-Modal Fulfillment