Dispatcher3 - D5.2: Verification and validation report

Abstract

The deliverable provides the results from the verification and validation activities within Dispatcher3 project. The document reviews the internal and external validation activities that were carried out during the course of Dispatcher3 project according to the plan defined in D5.1. Dispatcher3 is organised in three layers: data acquisition and preparation, predictive layer, with machine learning models, and prospective models, with the integration of the individual machine learning models in an interactive Advice Generator and an estimator of rotation/reactionary delay. This deliverable presents the verification and validation activities performed on these three components. For the data acquisition and preparation layer the data-pipelines, including the transformation verification and validation activities are described. In the predictive layer both the models developed for the first release and their evolution for the final prototype are described and presented. Finally, for the prospective layer, the interactive interface with its functional requirements is presented and verified, while the reactionary delay model is described, and different scenarios evaluated for its validation. The deliverable also describes the different internal and external activities and meetings, workshops and dedicated online site visits that have been performed during the duration of the project. Finally, the document assesses the verification of the high-level system-wide requirements identified at the beginning of the project in D1.1 – Technical resources and problem definition, and the research questions identified in the Verification and validation plan (D5.1).

This deliverable is part of a project that has received funding from the Clean Sky Joint Undertaking under grant agreement No 886461 under European Union’s Horizon 2020 research and innovation programme.

Peer Reviewed

Postprint (published version)