Marketplace Logistic via Answer Set Programming
summary
Implements marketplace order allocation with Answer Set Programming, using ASP Chef prototype for fast encoding and visualization and a production run with the clingo Python API. Formulation minimizes shipping cost first, then number of warehouses, then warehouse–product pairs via weak constraints. Benchmarks on generated instances show the greedy baseline is immediate but often misses savings while ASP finds the true optimum (often zero shipping) and yields fewer warehouses/pairs; runtime is higher but remains acceptable for the platform’s fast/online, slower/interactive, and offline stages.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:27Key actors
- 3:40Problem setup
- 5:49Why ASP fits
- 7:54ASP Chef prototyping
- 10:46Live demo workflow
- 14:01Optimization encoding
- 16:07Warehouse visualization
- 18:38Multi-level optimization
- 22:31Implementation benchmarks
- 25:43Conclusion
- 26:08Q&A
related talks
27mDecomposition-based job-shop scheduling with constrained clustering
47mPanel: AI impact on E-commerce: Friend or a Foe?
56mPeople, Ideas, and the Path Ahead
30mFrom Constraints to Cognition: A Hybrid Framework for Adaptive, Explainable(…)
51mA.I. constraint optimization with OptaPlanner
31mFormalizing and Reasoning about Contracts between Agents
13mBlawx: User-friendly Goal-Directed Answer Set Programming for Rules as Code
32mKeynote: Everywhere commerce