𝗢𝗽𝘁𝗶𝗺𝗶𝘇𝗶𝗻𝗴 𝗠𝗮𝗿𝗶𝘁𝗶𝗺𝗲 𝗣𝗼𝗿𝘁 𝗖𝗼𝗻𝗴𝗲𝘀𝘁𝗶𝗼𝗻
Ports often use simple first-come-first-served rules. These rules fail when weather changes or equipment breaks. They do not account for high delay costs.
I built AnchorFlow-AI to solve this. It is a Python simulation engine. It uses a multi-agent system to manage berths, pilots, and tugboats.
The system uses a Prompt-Agent-Skill design. This breaks complex tasks into small, reusable blocks:
- Triage Agent: Sorts the queue by delay cost and wait time.
- Berth Allocator: Matches ships to docks based on size and depth.
- Transit Agent: Checks if pilots and tugs are ready.
- Operations Coordinator: Manages the whole process and handles disruptions.
I tested this against a standard scheduling model. I added random storms and mechanical failures to see how the system reacts.
The results:
- Vessel waiting times dropped by 16.4%.
- Demurrage costs fell by over $480,000 USD.
- The system recovers quickly from port disruptions.
This project shows how intelligent agents can handle logistics better than static rules. It moves from simple planning to active orchestration.
You can run this simulation locally on your laptop using Python.
Optional learning community: https://t.me/GyaanSetuAi