Abstract
<p>The aim of this paper is to present the <strong>objectives, research directions and first highlight results</strong> of the <strong>Smart4RES</strong> project, which was launched in November 2019, under the <strong>Horizon 2020</strong> Framework Programme. Smart4RES is a research project that aims to bring substantial performance improvements to the whole model and value chain in r<strong>enewable energy (RES) forecasting</strong>, with particular emphasis placed on optimizing <strong>synergies with storage and to support power system operation and participation in electricity markets</strong>. For that, it concentrates on a number of disruptive proposals to support ambitious objectives for the future of renewable energy forecasting. This is thought of in a context with steady increase in the quantity of data being collected and computational capabilities. And, this comes in combination with recent advances in <strong>data science</strong> and approaches to <strong>meteorological forecasting</strong>. Smart4RES concentrates on novel developments towards <strong>very high-resolution and dedicated weather forecasting solutions</strong>. It makes <strong>optimal use of varied and distributed sources of data</strong> e.g. remote sensing (sky imagers, satellites, etc), power and meteorological measurements, as well as high-resolution weather forecasts, to yield high-quality and seamless approaches to renewable energy forecasting. The project accommodates the fact that all these sources of data are distributed geographically and in terms of ownership, with current restrictions preventing sharing. Novel alternative approaches are to be developed and evaluated to reach optimal forecast accuracy in that context, including <strong>distributed and privacy-preserving learning and forecasting methods</strong>, as well as the advent of platform-enabled <strong>data-markets</strong>, with associated pricing strategies. Smart4RES places a strong emphasis on <strong>maximizing the value from the use of forecasts in applications</strong> through advanced decision making and optimization approaches. This also goes through approaches to streamline the definition of new forecasting products balancing the complexity of forecast information and the need of forecast users. Focus is on developing models for applications involving storage, the provision of ancillary services, as well as market participation.</p>

Abstract
The generalized Schrodinger-Newton system of equations with both local and nonlocal nonlinearities is widely used to describe light propagating in nonlinear media under the paraxial approximation. However, its use is not limited to optical systems and can be found to describe a plethora of different physical phenomena, for example, dark matter or alternative theories for gravity. Thus, the numerical solvers developed for studying light propagating under this model can be adapted to address these other phenomena. Indeed, in this work we report the development of a solver for the HiLight simulations platform based on GPGPU supercomputing and the required adaptations for this solver to be used to test the impact of new extensions of the Theory of General Relativity in the dynamics of the systems. In this work we shall analyze theories with nonminimal coupling between curvature and matter. This approach in the study of these new models offers a quick way to validate them since their analytical analysis is difficult. The simulation module, its performance, and some preliminary tests are presented in this paper.

Abstract
The study of coastal boulder accumulations generated by extreme marine events, and of the energy and frequency involved in boulder transport, is of paramount importance in understanding the risk associated with extreme marine inundations. One of the frequently asked questions is whether the deposits are storm or tsunami-related, both events being characterized by different return periods. Boulder transport by storms was monitored on the west coast of Portugal. Significant changes were detected in boulders' position as a result of extreme inundation by the 2013/2014 winter storms. Results presented in this work indicate that the wave power associated with the "Christina" and "Nadja" storms occur once every three years. However, this interval is not supported by field observations of boulder displacement, which suggests that wave power over-predicts boulder movement in the study area. Furthermore, wave parameters from the "Christina" and "Nadja" storms were very similar, but have generated different impacts in the boulder accumulation described herein. Differences include the magnitude and direction of boulder movement, and are most likely associated with distinct tidal levels during the events. Higher tide levels generated an increase in the sea surface level and thus in the reach of waves, which generated displacement of larger boulders and consequent cross-shore contribution in boulder transport. Regardless, the combination of monitoring campaigns, wave data, and statistical modelling of extreme values indicate that boulder transport by storms is more frequent than initially expected. Based on recorded boulder movements, we present a conceptual model for boulder ridge formation and development and identify significant control of incoming flow by local geomorphological/topographical features. Storm events, not less frequent tsunamis, are identified as the events responsible for modulating this rocky coastline. These results question a direct attribution of coastal boulder deposits to tsunamis in coastal regions with a high risk of tsunami inundation.

Abstract
Given the plethora of Reinforcement Learning algorithms available in the literature, it can prove challenging to decide on the most appropriate one to use in order to solve a given Reinforcement Learning task. This work presents a benchmark study on the performance of several Reinforcement Learning algorithms for discrete learning environments. The study includes several deep as well as non-deep learning algorithms, with special focus on the Deep Q-Network algorithm and its variants. Neural Fitted Q-Iteration, the predecessor of Deep Q-Network as well as Vanilla Policy Gradient and a planner were also included in this assessment in order to provide a wider range of comparison between different approaches and paradigms. Three learning environments were used in order to carry out the tests, including a 2D maze and two OpenAI Gym environments, namely a custom-built Foraging/Tagging environment and the CartPole environment.

Abstract
Many real-world robotic scenarios require performing task planning to decide courses of actions to be executed by (possibly heterogeneous) robots. A classical centralized planning approach has to find a solution inside a search space that contains every possible combination of robots and goals. This leads to inefficient solutions that do not scale well. Multi-Agent Planning (MAP) provides a new way to solve this kind of tasks efficiently. Previous works on MAP have proposed to factorize the problem to decrease the planning effort i.e. dividing the goals among the agents (robots). However, these techniques do not scale when the number of agents and goals grow. Also, in most real world scenarios with big maps, goals might not be reached by every robot so it has a computational cost associated. In this paper we propose a combination of robotics and planning techniques to alleviate and boost the computation of the goal assignment process. We use Actuation Maps (AMs). Given a map, AMs can determine the regions each agent can actuate on. Thus, specific information can be extracted to know which goals can be tackled by each agent, as well as cheaply estimating the cost of using each agent to achieve every goal. Experiments show that when information extracted from AMs is provided to a multi-agent planning algorithm, the goal assignment is significantly faster, speeding-up the planning process considerably. Experiments also show that this approach greatly outperforms classical centralized planning.

Abstract
This paper aims to explore and describe fundamental factors that influence the knowledge and use of Information and Communication Technologies (ICT), in particular of Educational Software (ES) as a tool, by teachers who teach mathematics in any portuguese teaching cycle. The adopted method has a descriptive and exploratory nature, focusing on a quantitative paradigm. The study participants were 96 teachers who teach mathematics, from various elementary and secondary schools, from different regions of Portugal, as well as from other countries where the portuguese curriculum is inserted. The questionnaire, the chi-square independence test and Cramer's V test were used as instruments. Data analysis was performed using SPSS (version 25) and Excel (Office 2016). The results suggest that the age, gender and length of service of mathematics teachers may be factors that influence the knowledge and use of Kahoot ES and that having training may be an essential condition for the use of Modellus and Scratch ESs. We did not find any relationship of dependence between having training and the use of the rule and compass ES by the respondents.