US Navy looks to speed up ship maintenance with AI-powered scheduling
Following a contract award from the US Navy to demonstrate scheduling capability for surface warship maintenance, Stottler Henke tells Harry Lye how AI-powered software can improve ship readiness and maintenance efficiency.
// USS Nimitz undergoing maintenance. Image: US Navy
US Navy has awarded a contract to AI software developer Stottler Henke to create ‘a critical chain and critical path scheduling capability’ to improve surface warship maintenance. Under the contract the company will demonstrate a system that can solve large-scale maintenance problems and deliver a seamless picture of surface ship maintenance requirements.
The navy used critical chain project management (CCPM) methodology to manage operations in the past and now seeks to pull in this approach to improve maintenance operations which would increase fleet readiness.
Increasing maintenance scheduling efficiency with AI
At the heart of Stottler Henke’s solutions is a platform called Aurora, which was developed to help NASA with mission-critical scheduling. Boeing also used the tool during the build process of the 787 Dreamliner. Aurora promises increased efficiency over traditional scheduling tools by using AI to build on knowledge of the task at hand.
Explaining how the system would benefit the navy, Stottler Henke president Richard Stottler tells us: “The Aurora scheduling system is the world’s leading intelligent scheduling software solution that uses advanced artificial intelligence to increase throughput. Over the last 25 years, an increasing number of organisations have chosen Aurora to manage their high-value operations. This is because Aurora generates schedules that are more efficient than those produced by any other system it has been compared to, including Microsoft Project and Primavera.
“As the US Navy pushes to expand its fleet and meet a number of operational requirements, readiness is becoming increasingly important in the ability to field ships.”
“By increasing the throughput of the navy’s ship maintenance operations, Aurora will enable more maintenance to be performed in a shorter amount of time using the same resources (labour, equipment, space, etc.), thereby lowering costs. The time and resources that are saved can be put towards the maintenance of another vessel. In addition, shorter turnarounds will enable ships to return to sea more quickly.”
As the US Navy pushes to expand its fleet and meet a number of operational requirements, readiness is becoming increasingly important in the ability to field ships. Improving maintenance efficiency and turnaround times is a critical aspect of this process.
What AI-powered scheduling software can do
Stottler says the system has proven its worth in a number of scenarios including at Boeing, where Aurora managed resources more effectively than Boeing’s own system that the aircraft manufacturer had built specifically for its operations.
He explains the advantages of an AI-enabled system: “Aurora solves complex scheduling problems more effectively by encoding and applying sophisticated scheduling knowledge and decision-making rules, along with complex constraints and resource requirements. Aurora encodes attributes of individual tasks, groups of tasks, resources, resource sets, and constraints.
“This knowledge-rich approach enables Aurora to combine human expertise with mathematical algorithms to generate superior schedules.”
“Aurora’s built-in and user-supplied decision rules produce better schedules by considering the values of these attributes at key scheduling decision points such as determining which task to schedule next, assigning the best time and resources to each task to optimise the overall schedule, and handling situations in which some resources are unavailable when the task is scheduled. This knowledge-rich approach enables Aurora to combine human expertise with mathematical algorithms to generate superior schedules.
“In addition, Aurora uses advanced techniques such as bottleneck avoidance that pre-analyses the resources required by all of the tasks, prior to scheduling them, to ensure that the most constrained resources are assigned to the tasks that have the greatest need for them, resulting in better schedules.”
Using experience from other military projects
Stottler Henke’s has previously worked with the US military alongside other companies to “develop artificial intelligence systems that solve hard problems that defy traditional approaches”, Stottler says.
Aurora provides the core of the US Air Force’s Managed Intelligent Deconfliction and Scheduling (MIDAS) system, which is used to rapidly schedule communications between the Air Force Satellite Control Network’s ground stations and the satellites they manage. This role in the past was carried by specialist schedulers but is now managed automatically by MIDAS.
Applying Aurora to ship maintenance, the company believes, will lead to AI-powered scheduling tools becoming commonplace in the future.
“The primary reason is that AI-based systems can provide superior schedules quickly, and the value of greater efficiency, is tremendous for complex, resource-intensive, mission-critical operations, Stottler says. “In addition, it is very important to retain the corporate knowledge that is held in the heads of various naval expert schedulers. As many of the experts retire, it is important to capture their knowledge in some sort of knowledge repository, in order to be leveraged by AI systems.”
“If all the ships take an extended amount of time to be maintained or upgraded, having more ships does not necessarily mean there are more ships on the water.”
In terms of results, Stottler says Aurora could improve maintenance times by at least 5%, by a ‘conservative’ estimate, as similar tools in similar complex environments have achieved better results.
If the demonstration proves successful, the plan is for the navy to roll out an Aurora-based system across naval shipyards, maintenance centres, and ship facilities following government assessment.
“Aurora will help fill the gap for increased readiness by increasing maintenance throughput, Stottler says. “That is, readiness is a direct function of how fast maintenance can be performed. For example, if all the ships take an extended amount of time to be maintained or upgraded, having more ships does not necessarily mean there are more ships on the water.
“If, on average, Aurora enables just one additional ship to be on the water, compared to a conventional scheduler, it is as if the US Navy received another ship without purchasing anything.”