Evolutionary Computation methods applied to Operational Control Centers

Abstract

During the execution of an operational plan, there is the likelihood of this plan being affected by some disruptions caused by unexpected events. The disruptions affect at least three dimensions that airline companies and the operational control centers must take into account which are pas- senger, crew and aircraft. Usually, a disruption is a state during which the current operation being executed is affected by a deviation (which is large enough to cause a change) from the original plan, and sometimes unfortunately it leads to an unfeasible plan. Examples of events that might cause disruptions are bad weather, threats or terrorist attacks and aircraft malfunctions. Disruption Management, can be defined as the process that starts after the deviation from the original plan is detected. After the disruption, the plan is changed and it will no longer be as close as it was from an optimal plan or it can even turn into an unfeasible plan. Either way there is a need to review the plan and try to minimize the impact caused by the disruption. MASDIMA is useful to help Airline Operation Control Centers finding a solution to disrup- tions during an operational plan, and in order to improve both computing time and the quality of solutions, there is a need to improve the system. This will translate in minimising the impact both in terms of costs or delays. To deal with that problem three agents will be implemented that will reflect, each one, different evolutionary computation algorithms (Particle Swarm Optimisation, Ant Colony Optimisation and Genetic Algorithms) and are related to the aircraft dimension of the problem. These agents will be implemented on a Multi-Agent System named MASDIMA that represents an Operation Control Center.