Abstract
Critical health events represent a relevant cause of mortality in intensive care units of hospitals, and their timely prediction has been gaining increasing attention. This problem is an instance of the more general predictive task of early anomaly detection in time series data. One of the most common approaches to solve this problem is to use standard classification methods. In this paper we propose a novel method that uses a layered learning architecture to solve early anomaly detection problems. One key contribution of our work is the idea of pre-conditional events, which denote arbitrary but computable relaxed versions of the event of interest. We leverage this idea to break the original problem into two layers, which we hypothesize are easier to solve. Focusing on critical health episodes, the results suggest that the proposed approach is advantageous relative to state of the art approaches for early anomaly detection. Although we focus on a particular case study, the proposed method is generalizable to other domains.

Abstract
Global climate change is currently a focus issue because of its impacts on the most diverse natural systems and, consequently, the development of humanity. The electricity sector is a major contributor to climate change because of its long-standing dependence on fossil fuels. However, the energy paradigm is changing, and renewable sources tend to play an increasingly important role in the energy mix in Portugal. Due to the strong relationship between renewable energies and climate-related natural resources, the climate change phenomenon could have considerable effects on the electricity sector. This paper analyzes the effects of climate change on the energy mix in Portugal in the medium / long term (up to 2050). The proposed methodology is based on the simulation of climate scenarios and projections of installed power by type and consumption. The combinations of these conditions are inputted to an energy accounting simulation tool, able to combine all information and provide a characterization of the system state for each case. The most favorable forecasted scenarios indicate that a fully renewable electricity system is achievable in the medium term, in line with the objectives of the European Union, as long as investments in renewable sources continue to be stimulated in the coming years.

Abstract
This paper shows that the recent approach to quantitative typing systems for programming languages can be extended to pattern matching features. Indeed, we define two resource-aware type systems, named U and E , for a λ-calculus equipped with pairs for both patterns and terms. Our typing systems borrow some basic ideas from [19], which characterises (head) normalisation in a qualitative way, in the sense that typability and normalisation coincide. But, in contrast to [19], our systems also provide quantitative information about the dynamics of the calculus. Indeed, system U provides upper bounds for the length of (head) normalisation sequences plus the size of their corresponding normal forms, while system E , which can be seen as a refinement of system U , produces exact bounds for each of them. This is achieved by means of a non-idempotent intersection type system equipped with different technical tools. First of all, we use product types to type pairs instead of the disjoint unions in [19], which turn out to be an essential quantitative tool because they remove the confusion between âbeing a pairâ? and âbeing duplicableâ?. Secondly, typing sequents in system E are decorated with tuples of integers, which provide quantitative information about normalisation sequences, notably time (cf. length) and space (cf. size). Moreover, the time resource information is remarkably refined, because it discriminates between different kinds of reduction steps performed during evaluation, so that beta, substitution and matching steps are counted separately. Another key tool of system E is that the type system distinguishes between consuming (contributing to time) and persistent (contributing to space) constructors.

Abstract
This article serves to show how things are changing when it comes to business and social networking. Nowadays, it is hard to find a business that does not have an account on any social network, and we can safely say that social media is a crucial aspect for any business enterprise - to sell their products, to be seen and, obviously, to make more money. We talk about the possibility to mix innovation, business strategy and social media. To complement this research, we focused our work on a Portuguese start-up - Strain - that intends to prove marketing is changing and that it actually welcomes the change, with social networks at the base of their business. They intend to use online influencers to promote the image of a brand, free of charge, and earn discounts on it. It is, indeed, a win-win situation, where each of the three parts (the company itself, clients and Strain) is a winner one way or the other. © 2019 AISTI.

Abstract
Laser Guide Star [LGS] wave-front sensing is a key element of the Laser Tomographic AO system and mainly drives the final performance of any ground based high resolution instrument. In that framework, HARMONI the first light spectro-imager of the ELT [1,2], will use 6 Laser focused around 90km(@Zenith) with a circular geometry in order to sense, reconstruct and correct for the turbulence volume located above the telescope. LGS wave-front sensing suffers from several well-known limitations [3] which are exacerbated by the giant size of the Extremely Large Telescopes. In that context, the presentation is threefold: (1) we will describe, quantify and analyse the various effects (bias and noise) induced by the LGS WFS in the context of ELT. Among other points, we will focus on the spurious low order signal generated by the spatially and temporally variable sodium layer. (2) we will propose a global design trade-off for the LGS WFS and Tomographic reconstruction process in the HARMONI context. We will show that, under strong technical constraints (especially concerning the detectors characteristics), a mix of opto-mechanic and numerical optimisations will allow to get rid of WFS bias induce by spot elongation without degrading the ultimate system performance (3) beyond HARMONI baseline, we will briefly present alternative strategies (from components, concepts and algorithms point of view) that could solve the LGS spot elongation issues at lower costs and better robustness.

Abstract
In this paper, a mathematical model is developed to tackle a Consistent Vehicle Routing Problem, which considers customers with multiple daily deliveries and different service level agreements such as time windows, and release dates. In order to solve this problem, an instance size reduction algorithm and a mathematical programming based decomposition approach are developed. This solution approach is benchmarked against a commercial solver. Results indicate that the method solves instances of large size, enabling its application to real-life scenarios. A case study in a pharmaceutical distribution company is analyzed. Consistent routes are planned for several warehouses, comprising hundreds of orders. A simulation model evaluates the performance of the generated route plans. Significant improvements in terms of the total distance traveled and the total travel times are obtained when compared to the company's current planning process.