Design-Time Learning for Operational Planning Improvement

This document describes how the DESEREC framework can be enhanced through the usage of an offline learning loop. This mechanism consists basically of computational methods and techniques which enable the framework to gradually improve the effectiveness of its behaviour, by analysing internal data produced by some of the modules, applying some learning algorithms to them and to the information received from the administrator and using then the results for enhancing the framework operation. The learning loop presented here is strictly related to the improving the quality of the generated operational configurations, which is achieved by adjusting the decisions made by the Planning module while carrying out its task.