The floor of an ALC is like a well-orchestrated dance. However, when an unexpected failure occurs, it can introduce significant delays in not only the specific aircraft"s maintenance, but of the entire network of systems. 21st Century Systems, Inc. (21CSi) is in the position to answer this challenge with a concept called Reactive Aircraft Maintenance System via Evolutionary Scheduling (RAMSES). This project"s namesake was known for his organizing ability and capacity to move large projects along. Just as Ramses utilized the best technology of his day, RAMSES uses powerful algorithms to optimize the maintenance network in the face of unforeseen failures. The base optimization engine is an Evolutionary Algorithm (EA), which provides the ability to operate under the uncertainty an unforeseen failure creates through the use of Subjective Logic. This enables the algorithm to optimally select a maintenance plan that deals directly with the materials, personnel, and facility, yet allow the algorithm to handle the uncertainty of the availability of each. The result is a shop floor capable of effectively handling any unforeseen event. And, with 21CSi"s sustained superlative commercialization index, we are the right company at the right time to improve ALC operations.
Benefit: The RAMSES concept will have direct benefit to both the military aircraft depot as well as commercial aircraft depots. RAMSES provides a method to optimally reschedule the entire depot if an unforeseen event occurs that may cause major disruptions in the work flow. Behind the scenes, RAMSES is able to interlink all the components and activities at the depot using an evidential reasoning technology, the Evidential Reasoning Network (ERN). ERN technology allows us to describe the effects an unforeseen event would have on the components of the depot in terms of added or reduced delay. By incorporating uncertainty directly into the calculations, we are able to form conditional bounds on the effect. The EA then solves the optimization problem using these bounds. Therefore, when an unexpected event occurs, RAMSES sends out the rescheduling plan directly to the same system that the technicians are currently using. RAMSES will even develop contingency plans in the case of compound or cascading events. As far as the technician is concerned, the result is simply a change in the schedule. However, the depot supervisor, overseeing the entire operation, will notice how RAMSES maneuvers the depot floor to account for the unexpected event whilst maintaining the highest throughput possible given the anomaly. We also see this technology expanding beyond aircraft. Almost all fleet type operations have some form of maintenance shop. Land, sea, and air all require their vehicles to be periodically serviced to maintain the performance necessary for operational success. RAMSES is able to optimize scheduling from the smallest to the largest operations. The innovative ERN method to describe the relationships of the components and vehicles in the shop can be scaled for any size operation. The result is an integrated tool that saves time in a situation that usually only costs time and resources.
Keywords: Dynamic Re-planning, High Velocity Maintenance, Agent-based software, Evolutionary Algorithms, Subjective Logic, Evidential Reasoning, Uncertainty Handling, Contingency Planni
