Proactive Supply Chain Disruption Intelligence System

Predict. Simulate. Optimize.

An AI-powered decision support system designed to anticipate and mitigate supply chain disruptions at the port level through predictive analytics, simulation, and intelligent recommendations.

Overview

Global supply chains are increasingly vulnerable to disruptions caused by weather, congestion, and operational inefficiencies. Most existing systems are reactive, responding only after issues arise.

This project introduces a proactive intelligence pipeline that:

• Predicts congestion risk in advance
• Simulates operational impact
• Recommends optimal routing decisions

Key Features

Disruption Risk Prediction

Forecasts congestion probability and risk levels using ML models.

T+2 Forecasting

Predicts future congestion and berth occupancy.

Simulation Engine

Estimates delay, queue size, and operational impact.

Adaptive Triggering Logic

Runs simulation only when risk crosses a threshold.

Route Recommendation System

Suggests optimal routes based on cost, delay, and risk.

End-to-End Pipeline

Integrated flow: Prediction -> Simulation -> Recommendation.

System Architecture

User (Frontend) | v Frontend (Vercel) | v Backend API (FastAPI - Render) | v Intelligence Layer - Prediction Model (XGBoost) - Simulation Engine (SimPy) - Solution Engine (Gemma AI) | v Data Layer (PostgreSQL / Firebase) | v External AI (Google Gemma via OpenRouter)

Process Flow

User Input | v Prediction Model | v Risk Evaluation (Threshold) | | v v Low Risk High Risk | | v v Output Simulation Engine | v Solution Engine | v Final Output

Tech Stack

Backend and APIs

• FastAPI (Python)
• Node.js (Express)

AI and Analytics

• XGBoost
• scikit-learn
• NumPy, Pandas

Simulation

• SimPy

Data Layer

• PostgreSQL
• Firebase (Auth and storage)

Frontend

• HTML, CSS, JavaScript

Infrastructure

• Docker and Docker Compose
• Nginx (Reverse Proxy)
• Vercel (Frontend Hosting)
• Render (Backend Deployment)

Google Technologies Used

• Firebase: authentication and real-time data storage
• Gemma (Google AI): intelligent recommendations and solution generation

Deployment

• Frontend: Vercel
• Backend: Render (Dockerized FastAPI service)
• Database: PostgreSQL
• Auth and Data: Firebase

API Endpoints

Endpoint	Description
/run_pipeline	Full pipeline (prediction + simulation + recommendation)
/predict	Congestion prediction
/simulate	Delay and impact simulation
/recommend	Route recommendations

Example Output

{
	"prediction": {
		"congestion_probability": 0.92,
		"risk_level": "HIGH"
	},
	"simulation": {
		"estimated_delay_hours": 12.5,
		"queue_size": 8
	},
	"recommendation": [
		{
			"route": "Route B",
			"cost": 1200,
			"delay_factor": 6.2
		}
	]
}

Use Cases

• Logistics planning and optimization
• Port authority decision support
• Supply chain risk management
• Real-time disruption analysis

Future Enhancements

• Real-time data integration (weather, vessel tracking)
• Multi-port global network simulation
• Advanced time-series and deep learning models
• Scenario-based "what-if" analysis dashboard
• Full SaaS deployment