Abstract
Tackling uncertainty is becoming increasingly relevant for decision-support across fields due to its critical impact on real-world problems. Uncertainty is often modelled using scenarios, which are combinations of possible outcomes of the uncertain parameters in a problem. Alongside expected value methods, decisions under uncertainty may also be tackled using methods that do not rely on probability distributions and model different decision-maker risk profiles. Scenarios are at the core of these approaches. Therefore, we propose a scenario generation methodology that seizes the structure and concepts of genetic algorithms. This methodology aims to obtain a diverse set of scenarios, evolving a scenario population with a diversity goal. Diversity is here expressed as the difference in the impact that scenarios have on the value of potential solutions to the problem. Moreover, this method does not require a priori knowledge of probability distributions or statistical moments of uncertain parameters, as it is based on their range. We adapt the available code for Biased-Random Key Genetic Algorithms to apply the methodology to a packing problem under demand uncertainty as a proof of concept, also extending its use to a multiobjective setting. We make available these code adaptations to allow the straightforward application of this scenario generation method to other problems. With this, the decision-maker obtains scenarios with a distinct impact on potential solutions, enabling the use of different criteria based on their profile and preferences.

Abstract
Digitalization has spread across business and supply chains, becoming irreversible and affecting how companies run their businesses and fulfill their demand. This paper discusses the main aspects that propel and hinder digitalization in supply chains are. One could divide the boosters into two groups: the application of technological advances and circular boosters. On the other hand, the barriers are either sporadic or persistent. Despite the perceived barriers, if correctly applied, digitalization brings more benefits than problems to supply chains. Furthermore, recognizing this might help practitioners who are still reluctant about digitalization.

Abstract
In highly regulated areas such as healthcare there is a demand for explainable and trustworthy systems that are capable of providing some sort of foundation or logical reasoning to their functionality. Therefore, deep learning applications associated with such industry are increasingly required by this sense of accountability regarding their production value. Additionally, it is of utter importance to take advantage of all possible data resources, in order to achieve a greater amount of efficiency respecting such intelligent frameworks, while maintaining a realistic medical scenario. As a way to explore this issue, we propose two models trained with information retained in chest radiographs and regularized by the associated medical reports. We argue that the knowledge extracted from the free-radiology text, in a multimodal training context, promotes more coherence, leading to better decisions and interpretability saliency maps. Our proposed approach demonstrated to be more robust than their baseline counterparts, showing better classification performances, and also ensuring more concise, consistent and less dispersed saliency maps. Our proof-of-concept experiments were done using the publicly available multimodal radiology dataset MIMIC-CXR that contains a myriad of chest X-rays and its correspondent free-text reports.

Abstract
This article presents Pardinus, an extension of the popular Kodkod relational model finder with linear temporal logic (including past operators), to simplify the analysis of dynamic systems. Pardinus includes a SAT-based bounded-model checking engine and an SMV-based complete model checking engine, both allowing iteration through the different instances (or counter-examples) of a specification. It also supports a decomposed parallel analysis strategy that improves the efficiency of both analysis engines on commodity multi-core machines.

Abstract
Label noise is omnipresent in the annotations process and has an impact on supervised learning algorithms. This work focuses on the impact of label noise on the performance of learning models by examining the effect of random and class-dependent label noise on a binary classification task: quality assessment for photoplethysmography (PPG). PPG signal is used to detect physiological changes and its quality can have a significant impact on the subsequent tasks, which makes PPG quality assessment a particularly good target for examining the impact of label noise in the field of biomedicine. Random and class-dependent label noise was introduced separately into the training set to emulate the errors associated with fatigue and bias in labeling data samples. We also tested different representations of the PPG, including features defined by domain experts, 1D raw signal and 2D image. Three different classifiers are tested on the noisy training data, including support vector machine (SVM), XGBoost, 1D Resnet and 2D Resnet, which handle three representations, respectively. The results showed that the two deep learning models were more robust than the two traditional machine learning models for both the random and class-dependent label noise. From the representation perspective, the 2D image shows better robustness compared to the 1D raw signal. The logits from three classifiers are also analyzed, the predicted probabilities intend to be more dispersed when more label noise is introduced. From this work, we investigated various factors related to label noise, including representations, label noise type, and data imbalance, which can be a good guidebook for designing more robust methods for label noise in future work.

Abstract
Production scheduling generates a direct impact on several aspects of manufacture, such as the number of delays in delivery to customers, total flow time, as well as the percentage of equipment used. It must, therefore, constitute a priority in production management, which should seek to implement scheduling techniques that will lead to positive results from the perspective of the quality of the solution. However, the methodology cannot overlook the functional aspect of the time which has elapsed until the solution is reached. This study is based on a real and specific module software improvement into a company devoted to the development of ERP software systems (Enterprise Resource Planning). It presents a solution for the production scheduling module focused on flow-shop operations, comprising a total of nine dispatching rules. An additional solution for scheduling is also proposed, which resorts to metaheuristic simulated annealing. Both solutions are compared to each other by using the quality-functionality binomial approach. These two environments are further contrasted with a third, where no effective solution for production scheduling exists. The environment which includes scheduling through dispatching rules was compared to the environment where no production scheduling was implemented. The results obtained from this analysis show an improvement of 13%. The simulated annealing solution presents an improvement of 3,6% when compared to a solution which uses dispatching rules. This improvement implies one extra minute in the calculation of the final solution.