When talking about optimisation in bulk shipping, most people imagine a black-box optimisation algorithm that is magically capable of generating an optimal schedule for you. I don’t believe this will be a reality anytime in the foreseeable future. A massive amount of data points and their complex interconnectedness would need to be modelled into the code. For this we would need an army of coders that understand the business logic, processes and complexity. And even then, it would be extremely hard to validate its logic for errors.

What then, is a better approach? What I think we need is a toolset which visualises the data in a way that’s easily understandable, and that highlights inefficiencies, risks and issues which need to be addressed. Assistance algorithms could propose comprehensible and verifiable changes to the schedule, which are then evaluated and accepted by the planner. Then, slowly, based on the planners’ input, the system would learn and take on more and more tasks, giving planners the chance to focus on more complex problems and inefficiencies.

On a high level, there are 3 ways to approach fleet TCE optimised schedule building:

1. Build a full schedule manually & then in a step by step process seek the biggest yielding parcel or voyage level swaps/changes to improve fleet TCE

2. Build the schedule through a semi assisted process, where you’re commonly picking from a list of the highest fleet TCE contributing voyages available & then seek the biggest yielding parcel or voyage level swaps/changes to further improve fleet TCE 

3. Use heuristic auto-plan functionality to generate entire schedules using a restrictions rule engine, and then manually fix (& define) all the problems that the algorithm wasn’t aware of

Whatever the approach you choose, it’s important to be able to create a few variations, and then compare them to draw the next set of conclusions. In an optimal scenario, you would create versions of each of the above 3 types and then compare. For this to be meaningful there also needs to be a way to quantify the schedule risk, and then take that into account in the form of comparison KPIs.

This is a powerful approach for fixed cargo planning, but can dramatically increase efficiency when also used for narrowing down market cargo and cargo swapping options between shipowners.

How is your company optimising the fleet at the moment?