A formula for making better decisions

Last updated: 9 September 2019

It is a particularly nerve-jangling job. And the stakes are high too.

Our Weipa bauxite operation in Queensland, Australia, ships more than 30 million tonnes of bauxite every year.

Up to ten big ships come into the Weipa port each week destined to carry bauxite to customers in eastern Australia, Asia and beyond. Ships without a berth can accrue daily penalties of up to A$100,000, upset bauxite production at the mine feeding the port, and delay delivery to customers. Getting the ships in, loaded up and out on time and on budget means juggling tides, times, customers, the weather and 30 separate decisions per ship worth millions of dollars a year.

This is the job of Michelle Sjoberg, port scheduler.

"We would spend three to four hours every morning running scenarios manually in response to the latest changes to the variables – weather, arrival times and stock availability, and our experience."

That was fine, but there had to be a better way.

And there was. Rio Tinto hired industrial mathematicians to produce software that helps port schedulers like Michelle to make better decisions, more consistently and more often. The result? A more efficient, lower-cost shipping operation.

"We run those same scenarios and make evidence-based decisions in maybe 30 minutes. Plus we know what we're doing is backed by data, not opinion," says Rod Smith, chief adviser for planning and scheduling in Rio Tinto's Productivity & Technical Support team.

And the benefits are not limited to ports.

We have found many other opportunities where "mathematical optimisation" (that's the industry term) can help us accurately evaluate alternatives – fast. From when we first make contact with ore in the ground to getting finished products to our customers, our teams can respond to real-world problems and opportunities as they happen.