Sailboats have captivated me, as I suppose they have done for many participants in the maritime business; my office walls proudly display the requisite Americas Cup litho’s, as well as some good British renderings of J-Class boats, like “Endeavour Defeating Velsheda near the Needles”- evoking a mid- 1930s competition. Fast forward eight or nine decades, with the 36th Americas Cup now history- and planners already getting ready for the 37th set of races. Doing a little bit of research on some of the boats (nerdy hobby, I know), I discovered that the defending champion- the winner of the races held earlier this year, Emirates Team New Zealand (which apparently had racked up a slew of previous wins), had benefited from a “BOT”, created by the consultancy McKinsey. Importantly, the software tool utilized Artificial Intelligence (AI) to help steer the boat- figuratively, of course.
The bots were actually running a simulator that was testing the best design for the boats- which are really flying hydrofoils with sails at this point. As the consultants explained, the yacht designs were optimized by bots playing on the simulators and improving the boats’ contours…the bots key to success was “…the ability to take the sailors’ schedules out of the equation and test designs 24/7 on rapid repeat.” Unlike the humans who need some rest every now and then, the bots never have to go to bed and just keep improving. The name for the technology is “reinforcement learning.”
This process is not quite the same as robots actually driving ships- which is what I thought I was going to learn about when I stumbled upon some articles about the bots. There have been efforts to reducing crews, and automating some tasks, on deepsea vessels, with shoreside information systems having the ability to lend a hand, in what the International Maritime Organization calls “Degree Two” of vessel autonomy. Further degrees of automation, Three and Four, fully remote control and then onboard fully automated, are mainly conceptual at this point. These days, reducing crew sizes is not the primary concern (though I think it will happen). Instead, optimizing vessel configurations is now front and center.
The design of vessels, which now must conform to infinitely complicated constraints (think about EEXI and EEDI) or engage in trades which are optimal for their particular vessel type (think CII, AER and even EEOI) may also be the province of bots. Quite simply, all the permutations of technical, operational and trading combinations that need to be synthesized is way too overwhelming for humans to reckon with. In particular, the ability to superimpose a trading environment (with particular impacts on the CII stuff) on top of more technical vessel design aspects, would be something of a break-through. These things are not static; instead, they are subject to downward sloping “trajectories” for greenhouse gas emissions.
“The race’s outcome proves that reinforcement learning can be a transformational tool for process design, with potential applications across industries,” says the consultant- talking about the New Zealand team winning yachting’s prize. With all its criss-crossing variables, and downward sloping glide paths, the new tools maybe just the ticket for the increasingly complicated problems facing commercial shipping. Carbon intensity, for shipping, is closely tied to trades and cargo- dependent, in turn, on the tons and mileage of commodity flows (always the stuff of simulations and “what ifs”). The intensity is still a concept that needs very much to be homogenized with the design side, which has entered a true experimentation phase. The latter sounds a little bit like the great changes as yacht design has evolved. Talking about reducing resistance at the waterline!!!
If the bots can assist in designing vessels that can actually navigate all the new constraints imposed by decarbonization, that would be an important advancement. So, those flying yachtsmen may have something to offer the more mundane world of commercial shipping. And those 1930’s era J-Boats that I love…well, my office wall loves them too.