Fairness analysis with cost impact for Brasilia's Flight Information Region using reinforcement learning approach

Abstract

To analyze fairness between passengers and airlines considering financial cost, the management of the adaptation for air traffic flow in a heuristic and dynamically manner is studied in this research. Multi-Agent theory with reinforcement learning approach is used as a basic methodology integrated with a system of Decision Support System Applied to Tactical Air Traffic Flow Management (SISCONFLUX). The objective to develop this model is to increase the safety preserve and reduce the air traffic congestions. Reward structure with evaluation functions of financial cost and delay's impact is proposed for related flights using real data from Brasilia's Flight Information Region (FIR-BS). With the developed model, the experimental results show that the time delay is 25% less than the results computed only by Graph Theory with the same data, and fairness considering financial cost factor can be used together with congestion scenario in the air traffic management without affecting safety and flow factors.