Autonomous Replenishment Agent Stress Simulation Builder

Role: Staff Supply Chain Automation Architect evaluating autonomous retail inventory and replenishment agent workflows.

Context

• Retail Merchandising Sector: {{retail_category}}
• Agent Decision Module: {{inventory_agent_logic}}
• Baseline Supply Shock Dataset: {{historical_stockout_data}}
• Lead Time Uncertainty Range: {{supplier_lead_time_variance}}
• Cost Overrun Threshold: {{cost_discrepancy_cap}}
• Evaluation Horizon: {{evaluation_window_days}}

Task

Produce an automated Python simulation script that feeds synthetic supply disruptions, surge demand, and lead-time delays into {{inventory_agent_logic}} over {{evaluation_window_days}} days, measuring fill-rate preservation and penalizing order cost spikes above {{cost_discrepancy_cap}}.

Method

1. Model synthetic daily demand curves tailored to {{retail_category}}, incorporating weekend peaks and seasonal stockout risks from {{historical_stockout_data}}.
2. Inject stochastic lead time delays governed by {{supplier_lead_time_variance}} to test the agent's buffer calculation.
3. Execute step-by-step stateful simulations where {{inventory_agent_logic}} receives current stock levels, open purchase orders, and forecasts.
4. Compute inventory holding costs, stockout penalties, and purchase cost deviations against the static threshold {{cost_discrepancy_cap}}.
5. Evaluate autonomous reorder quantity decisions against ground-truth dynamic safety stock formulas.
6. Log automated pass/fail flags for bullwhip amplification, phantom inventory recovery, and minimum order quantity violations.
7. Output tabular performance diagnostics and export scenario results into structured CSV/JSON formats.

Constraints

• The script MUST utilize standard data science libraries (numpy, pandas) and provide fully reproducible seeds.
• The evaluation MUST NOT permit negative inventory states without triggering severe penalty flags.
• You MUST generate mock interfaces for {{inventory_agent_logic}} so the script runs out of the box.
• All metrics must strictly evaluate decisions across the full {{evaluation_window_days}} timeline.

Output format

• Section 1: Simulation configuration and test parameter constants.
• Section 2: Mock Agent Interface and Supply Environment Simulation Class.
• Section 3: Core Evaluation Runner and KPI Calculator (Stockout Rate, Cost Overrun, Bullwhip Index).
• Section 4: Main execution block running 3 distinct disruption scenarios.

Self-review

1. Confirm that {{supplier_lead_time_variance}} is mathematically applied to replenishment delivery steps.
2. Verify that budget overruns exceeding {{cost_discrepancy_cap}} trigger explicit assertion failures.
3. Check that the script executes cleanly within standard Python 3 environments without proprietary dependencies.