We built a road transport optimizer. Then we took it to sea.
Turns out, optimization isn’t specific to roads. If you know the time and distance between two points, the same logic applies. Whether those points are depots and delivery sites, or ports and LNG cargoes.
We recently had a conversation with a major energy company. They’re in the LNG business, and they have a scheduling problem. They need to decide between long-term sales agreements with customers and spot contracts where the price varies with the market. Their schedulers are trying to do this by hand. Spreadsheets. Manual guesswork.
They weren’t getting the results they wanted.
So we stripped the irrelevant stuff out of our road transport optimizer, put in nautical information and test data, and it worked fine. We’re now demo-ing a version that handles their specific problem.
The Problem We’re Solving
LNG shipping is complex. A fleet of vessels moves cargo between ports globally, and schedulers need to balance two competing demands. On one hand, there are long-term Sales & Purchase Agreements (SPAs), lower margin but obligated deliveries that keep the lights on. On the other, there are spot market cargos, high margin but opportunistic, and you only get them if you can deliver when the price is right.
Schedulers are trying to do this by hand. They’re using spreadsheets and other solvers that aren’t directly applicable to their problem. They needed something that could handle the complexity of a 30-vessel fleet, strict contractual time windows, and the trade-offs between different types of cargo.
What We Built
We started with our Dynamic Transport Optimiser (DTO) – a platform that’s already proven in road transport, delivering solutions that are up to 50% closer to optimum compared to traditional optimisers. The DTO handles complex routing and scheduling problems across industries including transport, manufacturing, forestry, and mining.
Then we adapted it for maritime logistics.
The model we developed supports:
- Fleet & Locations: A 30-vessel LNG fleet operating globally, with realistic port and ocean nodes.
- Contractual Demand: A mix of long-term SPA deliveries (around 75% of volume, heavily weighted towards Japan, South Korea, and China) and opportunistic spot market orders (around 25%).
- Economic Variables: Time-based charter rates, distance-based fuel consumption, fixed port fees, and cargo-specific revenues.
- Time Windows & Service Durations: Strict 72-hour windows for cargo pickup and delivery, plus fixed 24-hour service times for loading and discharging.
The core objective is straightforward: maximize the net profitability of the LNG fleet by strategically balancing those low-margin obligated deliveries against high-margin opportunistic spot cargos.
What’s Next
The demo is working. The solver handles the constraints properly, even with the complexity that comes with global shipping. Along the way, we realized the interface needed some love. So we’re building a new one; cleaner, more intuitive.
We’re also looking at future extensions. Split deliveries, where a single cargo is discharged across multiple ports. Tide windows that restrict when a vessel can safely berth. Load management constraints for partially filled LNG tanks .
This is a new space for us. Maritime. Energy logistics. We’re excited to see where it goes.
