Cellular neural networks and multicriteria optimization

This deliverable is the output from T5.1 and T5.2. As such, it reports on the developed self-supervised machine-learning predictor that is capable of adjusting its structure and parameters autonomously. The predictor is implemented as a CeNN, with templates from an echo state network, that can be trained for all the data-driven modelling tasks from BatCAT with a high expected estimation per-formance. Moreover, from T5.2, ITWM’s pre-existing MCO methodology and tools are integrated into the BatCAT architecture and deployed as an IIDSS module, including an in-depth documentation and training for developers and users.