Defense, Demand Planning

Enabling dynamic flight scheduling with network analysis and AI

December 2020

When it comes to logistics, scheduling is vital. Whether it is cars, trucks, ships or aircrafts moving materials, products or people, there are many variables and costs to consider when putting together a schedule that works to achieve objectives.

While many companies and other organizations may have a system that has worked in the past, those methods often become ineffective as growth occurs. Billion dollar companies, public universities and government agencies are largely building schedules the old-fashioned way using Excel spreadsheets or, on rare occasions, scheduling optimization software like IBM CPLEX. Both of these solutions are difficult to manage in especially complex, dynamic environments where the number of activities, resources and constraints are high.

Schedulers must manage a massive number of choices to be made in the face of complex constraints and competition for resources. Poor resource allocation choices and scheduling errors can create major strains that often come with hefty financial repercussions. Human schedulers, which are still the primary resource for scheduling in many organizations, are not often able to do an adequate job of evaluating which versions of the schedule are best for the organization’s strategic goals because they simply don’t have time to properly analyze all of the often billions of options. This means many organizations are utilizing schedules that are not built to maximize one or more key performance indicators (KPIs) such as throughput, cost minimization, quality, customer satisfaction and more. Outdated scheduling methods are part of the reason logistical scheduling issues are common, but they no longer have to be.

Artificial intelligence (AI) can be applied to help businesses gain operational and organizational improvements. It is now possible to apply specialized AI techniques to quickly and effectively produce a schedule purpose-built to achieve an organization’s specific goals. When you consider how many possible versions of a schedule may exist, often millions or even billions, the motivation for tapping into AI becomes very clear.

The Problem

The United States Air Force and Marine Corps, branches within the U.S. Department of Defense (DoD), were struggling with flight scheduling. Each branch was utilizing hundreds of thousands of personnel hours annually on scheduling flights and subsequently adjusting those schedules as issues and changes arose. The lack of scheduling efficiency was leading to the underutilization of resources and personnel who felt stressed and overworked. It also made it difficult to manage the impact of scheduling changes that were occurring rather frequently.

Both the Air Force and Marine Corps typically schedule certain flights months in advance of takeoff. Yet shifting budgets and priorities as well as maintenance issues would often lead to significant changes to the schedule. Some events, like test or training flights, have complex dependencies that regularly lead to cascading changes to the master schedule. The end result was a tangled mess for schedulers who had the daunting task of reorganizing the schedule after every change.

The Solution

Vertex Intelligence, a data science company, began working within the Air Force and Marine Corps to help alleviate the flight scheduling issues. The company developed a customized constraint model and data interface, and configured and integrated it into its existing cloud-based AI scheduling solution to generate a schedule which maximizes the number of weekly scheduled flights while balancing personnel constraints and aircraft maintenance schedules.

Data from personnel management systems, historic flight schedules and maintenance records were all analyzed and integrated into the solution which operates by utilizing algorithms inspired by natural selection to intelligently find solutions that satisfy all constraints and objectives. AI is used to intelligently search through all possibilities, and over time, can even learn from what has worked in previous schedules. With the new system in place, the human scheduling manager only needs to supply the data inputs and define the KPIs. The solution does all the number-crunching, and the manager gets more time back in the work day.

Results

Rather than costing potentially hundreds of hours of work from on-staff personnel, the dynamic flight scheduler from Vertex Intelligence builds schedules in just minutes, which has reduced the time spent manually scheduling by more than 80 percent on average. With the need for spreadsheets and manual processes eliminated, employee satisfaction has improved. Additionally, the schedule and scheduling process became adaptive to changes in priorities and maintenance work. This increased the reliability of the long-range schedule while keeping it nimble enough to evolve as-needed. And, thanks to AI, the system will only become more accurate as time goes on and it learns from previous scheduling scenarios. Finally, the solution has helped to reduce unplanned aircraft maintenance.

The new dynamic flight scheduling system has proven to be transformative to the Air Force and Marine Corps, and both entities look forward to continuing to utilize the AI solution from Vertex Intelligence in the future.

Let us solve your problem.
Manufacturing, Competitive Intelligence

Competitive Intelligence and Performance Optimization Models Provide Smarter Insights for Manufacturer Selling Products Online

How manufacturers can improve online retail sales performance with better competitive intelligence

Read the case study →

Construction, Predictive Modeling

Forecasting cash flow with predictive modeling

Accurate cash flow forecasting drives profitability in construction industry

Read the case study →

Defense, Demand Planning

Enabling dynamic flight scheduling with network analysis and AI

How military flight schedulers leverage AI to improve operational efficiency

Read the case study →

Contact us