Validated Inventory Optimization Framework for E-commerce: Data & Code
Description
This dataset contains the data, figures, and code accompanying our validated inventory optimization framework for e-commerce. The framework optimizes the cycle service level (CSL) under a standard (Q,R) policy, balances holding, ordering, and shortage costs, and reports a U-shaped total cost curve with a minimum around 95% CSL (fill rate ≈ 99.8%). We provide both “policy-stable” sensitivity (CSL fixed at 95%) and “re-optimized” sensitivity (CSL re-optimized under ±20% parameter shifts), as well as pooling analyses (constrained vs. unconstrained) and transshipment opportunities.
Files
Steps to reproduce
Steps to Reproduce 1. Download/unzip the dataset so it contains code/, data/, figures/. 2. Python 3.10+: python -m venv venv && source venv/bin/activate pip install numpy pandas scipy matplotlib 3. From the dataset root, run: python code/complete_framework_v2-3.py 4. Outputs: CSVs in data/ updated, figures in figures/ regenerated; key metrics written to data/framework_summary.json. 5. Quick check: optimal CSL ≈ 0.95, fill rate ≈ 0.998, min total cost ≈ 0.351M CNY, ABC targets 98/95/90.
Institutions
- University of Wisconsin Madison