SODA 2026 ES Let’s talk

Berth and loading-dock optimisationNovember 2026 edition

The berthing order that costs the least in port.

The optimal berthing order, explained with AI

  • Alongside
  • On demurrage
  • Under way or at anchor

Explanations written with AI from the optimal plan; the figures come from the optimisation model.

When ships queue in port, the berthing order decides what you pay in demurrage and in operating costs, because every berth has its own handling rate and cost. SODA 2026 computes the lowest-cost plan in seconds, on your own computer.

18%

lower demurrage and operating cost than first come, first served, in the case published in TOP (1995): 25 ships, 5 berths, 8 days. Optimal in 3.5 seconds.

See the published case

Against first come, first served

The same bay and the same 14 ships, with both plans side by side. Press Play or drag the timeline.

First come, first served
Demurrage
Operations
Total

Optimal plan
Demurrage
Operations
Total

Day 1, 00:00

  • Under way or at anchor
  • Alongside, working cargo
  • On demurrage: laytime expired
  • Anchorage

Teaching example with made-up data. J1 and J2 take all ships; J3, medium and small ones; J4, small ones only, because it has less draught. The faster the jetty, the more its hour of operation costs: $600, $450, $300 and $200 at J1, J2, J3 and J4. Demurrage runs from the end of laytime (40, 32 and 20 hours after arrival) at $45,000, $28,000 and $12,000 per day for large, medium and small ships. Under first come, first served, each ship goes to the jetty where it finishes earliest; the optimal plan minimises demurrage plus operating costs with an integer programming model of the same type as SODA 2026’s, solved with the free HiGHS solver in about a second.

Show both plans as a table

How it works

You load your ship calls; SODA 2026 does the rest.

Optimising with mathematical programming usually needs a specialist to write the model, and another call to that specialist whenever something changes. Not with SODA 2026.

  1. YouFill in an Excel template with your berths and expected ships: arrival, handling time and operating cost at each berth, laytime and demurrage rate.
  2. SODA 2026Writes the optimisation model and solves it: which ship goes to which berth and when, at the lowest demurrage and operating cost.
  3. YouGet the optimal plan in a local web app, with tables and reports showing the cost of each ship.

An arrival changes or a berth closes? Change it in Excel and solve again. No waiting for anyone.

The published case

What you save over first come, first served.

18.3%lower demurrage and operating cost than first come, first served
3.5 sto the optimum with one-hour periods, on a laptop
0.15 swith six-hour periods

It is the case I published in the journal TOP in 1995, with the ships anonymised: 25 ships, 5 berths and 8 days, solved today with SODA 2026 and the free HiGHS solver. The benchmark is strict first come, first served, in which each ship, on arrival, goes to the berth where it finishes earliest.

An experienced planner does better than that, so 18% is a ceiling for that episode, not a promise. In a quiet week there is nothing to optimise; after a storm, a great deal. That is why the diagnosis uses your own call history.

Parra, E. (1995). Demurrage costs optimization in ports using binary mathematical programming models. TOP, 3(1), 145–159. doi.org/10.1007/BF02574808

Speed

The optimum, in seconds.

The shorter the periods, the more precise the plan and the larger the model. This is how the published case (25 ships, 5 berths, 8 days) performs up to the proven optimum, on a four-second scale:

01234 s6 h and 4 h: 0.15 and 0.14 s2 h: 0.63 s1 h: 3.5 s
PeriodVariablesConstraintsTime
6 hours1,4841720.15 s
4 hours2,1942460.14 s
2 hours4,4504680.63 s
1 hour8,9019113.5 s

Solver time with the free HiGHS 1.15 solver on a Lenovo Yoga Slim 7 laptop (Intel Core Ultra 7 258V, 32 GB); median of three runs. The front-page example, 14 ships and 4 jetties with one-hour periods, solves in about a second.

Why SODA 2026

The optimum, with the costs you actually pay.

The optimum, not a rule
Of all possible berthing orders, the one with the lowest total cost. Not just a good solution: the optimum of the model, and a defensible one.
Demurrage and operations, berth by berth
Each ship’s demurrage according to its laytime and rate, plus operating costs, which differ from berth to berth: each has its own handling rate and hourly cost. It counts money, not hours.
Your terminal, as it is
Each ship only at the berths that suit it, different handling times at each berth and ships with fixed berthing times.
Replan in seconds
The published case solves to optimality in 3.5 seconds on a laptop, so you can replan every time an arrival changes.
On your own machine
It runs on your computer, with no cloud and no annual fees; your data never leaves the terminal. Free HiGHS solver by default; Gurobi if you need it.
AI on the results
Talk to the plan in Claude or ChatGPT with a prompt prepared with the data.

What is at stake

How much could your terminal save?

An order-of-magnitude estimate with your own figures. The optimal plan cuts demurrage and, by putting each ship at the berth that suits it best, operating costs too.

For example, 500 calls a year, 15% of them with one day of demurrage at $30,000 per day, add up to $2.25 million.

Estimated savings
On demurrage
On operations

Starting values are examples, not figures from any particular terminal. In the front-page example, the optimal plan cuts demurrage by 16% and operating costs by 3%. The real figure comes from the diagnosis, with your call history and your costs.

Request the diagnosis

Why now

Demurrage rates follow the freight market, and in 2026 tanker freight rates soared: every day of waiting costs more than it did a year ago. I explained why in The Conversation (in Spanish): «Cómo se fija el precio de llevar un barril de petróleo por mar» (October 2026).

Ask the plan

Talk to the optimal plan.

Copy a prompt that already contains the front-page example and both plans, paste it into Claude, ChatGPT or whichever assistant you use, and ask what you like: why a ship changes jetty, who loses out, or what would need recalculating if a berth closes.

Open Claude Open ChatGPT

Then carry on with these questions. Each one copies separately:

  • Why does the small V2 go to jetty J4 in the optimal plan, even though it takes longer there?
  • Which ships lose out in the optimal plan, and why is it worth it?
  • How much of the saving comes from demurrage and how much from operations?
  • If J2 went out of service from hour 30, what would need to be recalculated?
Show the full prompt
Act as the planning manager of a port terminal. You will analyse two berthing plans for the same congestion episode, computed with an optimisation model of the same type as SODA 2026's. The data are made up, for teaching purposes.

CONTEXT
- 14 ships and 4 jetties. Times in hours from the start of the episode (hour 0 = day 1, 00:00).
- Jetties: J1 and J2 take all ships; J3, medium and small ones; J4, small ones only.
- Handling time by ship type and jetty (h): Large, J1 30 and J2 34; Medium, J1 20, J2 22 and J3 26; Small, J1 10, J2 11, J3 12 and J4 14.
- Operating cost per hour alongside: J1 $600, J2 $450, J3 $300 and J4 $200. The faster the jetty, the more its hour costs.
- Demurrage: each ship has a laytime counted from its arrival; every hour it finishes after its laytime ends is paid at its rate. Large: 40 h of laytime and $45,000/day. Medium: 32 h and $28,000/day. Small: 20 h and $12,000/day.
- Plan A, "first come, first served": each ship, as it arrives, goes to the compatible jetty where it finishes earliest, regardless of cost.
- Plan B, "optimal plan": minimises demurrage plus operating cost; between plans of equal cost, it lets the ship that arrived first berth first.

DATA 1. Ships
SHIP;TYPE;ARRIVAL (h);LAYTIME ENDS (h)
V1;Medium;0;32
V2;Small;1;21
V3;Large;2;42
V4;Medium;3;35
V5;Medium;5;37
V6;Small;6;26
V7;Large;8;48
V8;Medium;10;42
V9;Small;12;32
V10;Large;16;56
V11;Medium;20;52
V12;Small;30;50
V13;Medium;44;76
V14;Small;60;80

DATA 2. Plans (start and finish in hours; demurrage in hours paid; cost = demurrage + operations, in $)
SHIP;A JETTY;A START;A FINISH;A DEMURRAGE;A COST;B JETTY;B START;B FINISH;B DEMURRAGE;B COST
V1;J1;0;20;0;12000;J1;0;20;0;12000
V2;J2;1;12;0;4950;J4;1;15;0;2800
V3;J2;12;46;4;22800;J2;2;36;0;15300
V4;J3;3;29;0;7800;J3;3;29;0;7800
V5;J1;20;40;3;15500;J3;29;55;18;28800
V6;J4;6;20;0;2800;J4;15;29;3;4300
V7;J1;40;70;22;59250;J1;20;50;2;21750
V8;J3;29;55;13;22967;J2;36;58;16;28567
V9;J4;20;34;2;3800;J4;29;43;11;8300
V10;J2;46;80;24;60300;J1;50;80;24;63000
V11;J3;55;81;29;41633;J3;55;81;29;41633
V12;J4;34;48;0;2800;J4;43;57;7;6300
V13;J1;70;90;14;28333;J2;58;80;4;14567
V14;J4;60;74;0;2800;J4;60;74;0;2800

TOTALS
Plan A: cost $287,733 = demurrage $163,583 (111 h, 8 ships) + operations $124,150.
Plan B: cost $257,917 = demurrage $137,417 (114 h, 9 ships) + operations $120,500.

HOW I WANT YOU TO WORK
1. Start with a five-line executive summary for the terminal manager.
2. Explain why plan B costs less even though it does not reduce demurrage hours, and how much of the saving comes from demurrage and how much from operations.
3. Point out the three berth changes that save the most, with figures taken from the data.
4. Then answer my questions using only these data. If something cannot be known from them, say so. If a question requires recalculating (closing a jetty, changing an arrival, a rate or an operating cost), do not make up the result: explain which scenario should be run in SODA 2026 and what you would expect to see.
Answer in English.

The AI reasons about the plan already computed; it does not recompute it. To evaluate a new scenario with a guaranteed optimum, that scenario is solved in SODA 2026.

Who it is for

Wherever ships queue to berth.

SODA 2026 pays off where there are few berths, irregular arrivals and a waiting cost that gets paid.

Liquid bulk terminals

Oil products, chemicals and biofuels: several jetties and customers who pay for every hour of demurrage.

LNG and LPG terminals

One or two berths, very expensive ships and agreed discharge windows.

Dry bulk

Grain, cement, fertilisers or ore, with demurrage and despatch, just as with tankers.

Industrial ports

Multipurpose public quays that the port authority shares among several operators.

Plants with their own terminal

Refineries, chemical plants and power stations that receive their raw materials by sea.

It works best where the quay is divided into defined berths. For continuous container quays there are specialised tools.

Beyond the port: loading docks.

A distribution centre is a port on land, with trucks instead of ships and loading docks instead of berths. In Spain, waiting has had a legally recognised cost since 2022: if a truck waits more than one hour, from the moment it is made available until its loading or unloading is completed, the carrier may claim compensation for detention.

Savings appear at peak hours, when docks are scarce; with docks to spare, first come, first served is already close to optimal.

Article 22 of Spanish Law 15/2009 of 11 November on contracts for the carriage of goods by land, as amended by Royal Decree-Law 3/2022 of 1 March.

Equivalences between the port and the distribution centre
In portIn a distribution centre
ShipTruck
Berth or jettyLoading dock
Arrival timeAppointment or actual arrival
Handling time at each berthLoading or unloading time at each dock
Berths suitable for each shipDocks suitable for each truck: reefer, ramp, dangerous goods
Laytime and demurrageFirst hour and detention

The same model applies to aircraft stands at an airport, loading tracks at a rail terminal or parallel machines in a workshop: jobs that arrive, stations that serve them and a cost of waiting.

How we work

Success-based, no fixed fee.

No savings, no invoice.

  1. Free diagnosisI review your operation for one day (berths, arrivals, contracts and demurrage paid) and tell you frankly whether SODA 2026 adds value. No cost, no commitment.
  2. Success feeUsing your call history, we compare what happened with the optimal plan for those same calls, assess the improvement over one year, and I charge a percentage of that first year’s savings. Savings are measured on costs you already know: the demurrage you paid and what it costs to operate each berth.
  3. Afterwards, tailored to youA software licence for your planning team and, if you wish, ongoing service.
Request a free diagnosis

About the author

Enrique Parra Iglesias

Telecommunications engineer and PhD in Economics. Tenured associate professor (profesor titular) at the University of Alcalá, where he has taught mathematics, operations research and energy economics.

Thirty-five years at the CEPSA group (now Moeve), from operations research analyst and project manager to CFO of Refining, with twenty years in charge of budgeting and management control and seats on several boards of directors. He has developed more than twenty software systems for management control and operations research; SODA is one of them, and its model was published in TOP in 1995.

If you would like the full picture, I will send you my CV.

Request the CV

Related publications

By the same author: SOT II 2026, optimisation of production, storage and transport.

Contact

Let’s talk about your terminal.

A free diagnosis, with no commitment, is the starting point.

Enrique Parra Iglesias
Tenured associate professor, University of Alcalá. Author of SODA 2026 and SOT II 2026
Email: eparra@io-e.com
CV available on request