Enhancing a Domain-specific Modeling Language for Runway Assignment Systems with experiental knowledge based Model Guidance Methods

Abstract

The efficient use of runway capacities continues to gain in significance in modern Air Traffic Management. Over the past years an increase in air traffic has been observed, which leads in conjunction with limited airport capacities to cost-intensive waiting times and an increased workload for air traffic controllers. Decision support systems should enable a strategic decision-making even in periods of high traffic and therefore an efficient degree of utilization of the runway capacities. The adaptation of a support system to new airports is associated so far with a considerable expenditure of work and the precondition of an adequate knowledge of the existing domain constraints. The development of a domain-specific modeling language (DSL) for the scope of arrival management enables a development at domain level and in combination with a code transformation on this basis a generic creation of airport-specific systems. Especially the generation of a valid and optimized runway assignment system is subject to a variety of constraints that have to be taken into account. To cope with constraint restrictivity and large solution spaces to a model, in this approach airport-specific data, legal requirements and the experiental knowledge of domain experts (air traffic controllers and developers of arrival management systems) is gathered and captured in a knowledge base. Validation of a model is carried out making an alignment between an actual model state and data and referring constraints in the knowledge base. Further constraint reasoning methods can be triggered to support the modeling process by deriving conceivable modeling steps to guide the modeler towards a valid model state. The developed modeling language AMAN-ML (Arrival Management Modeling Language) for the runway assignment component of an arrival management system will be exemplified in this paper including its composition and functionality. Furthermore the integration of factual and experiental knowledge into model guidance methods and the resulting modeling process will be introduced