- IBM has developed a deep-learning framework capable of ocean wave forecasting that is 12,000% faster than traditional physics-based models at comparable levels of accuracy, IBM reported recently.
- As a result, computational costs have been greatly reduced, meaning that access to the forecasting information can be accessed by the shipping industry, allowing for better voyage routing in inclement weather as well as ease in achieving a desired metric goal, like fuel consumption amount and voyage time.
- This AI-based tech is part of an ongoing trend for companies to use AI algorithms to calculate and predict different kinds of metrics for a company or an individual.
Wave forecasting can reduce risk and increase cost savings for the marine industry, but IBM's new tech can be applied to a variety of industries and companies.
Supply chains must always factor known and unknown risk into their supply chain plans. Many companies even create their own calendars outlining events such as the Chinese New Year and contract negotiating season to ensure they're shipping at the right time to save money. Weather is the true unknown, however, as evidenced by the fact that extreme weather has caused 29% more disruptions since 2012. To be truly prepared includes assessing suppliers' level of risk, especially those located along the usual paths of hurricanes, tornadoes or other major natural disasters.
Although ocean cargo carriers are now massive, they still remain vulnerable to heavy winds, surging waves, and strong storms. “Marine industries are subject to huge uncertainties,” Fearghal O’Donncha, IBM Researcher, told Supply Chain Dive. In fact, wave predictions may be just the tip of the iceberg when it comes to understanding weather patterns via deep learning methods.
“We chose waves largely because of the complexity. If we can do well with waves we can do well with most things," O'Donncha continued. “The more data you add, the better it learns. What we’re focusing on more moving forward is how we can work with other organizations to extend it from research to an operational component that can provide value to a number of industries, such as the supply chain industry."