Luís Guimarães

Abstract
Ensuring process quality in modern manufacturing is increasingly challenging due to the complexity of production processes and reliance on skilled operators, which can lead to suboptimal solutions and poor quality. To address these challenges, we introduce a novel, unsupervised, robust, nonparametric control chart for Phase II monitoring. This chart tracks the degradation of a quality characteristic using a condition index that captures mean and scale shifts without relying on assumptions, offering high flexibility and adaptability. Comparative studies with state-of-the-art nonparametric schemes demonstrate faster detection capabilities and competitive accuracy across various scenarios. We validate our approach through its application in the glass container production process, showcasing its effectiveness in monitoring multiple defective rates. Although tested on defective rates, the methodology is adaptable to any quantifiable quality characteristic.

Abstract
Uncertainty is pervasive in modern manufacturing settings. In order to cope with unexpected events, scheduling decisions are commonly taken resorting to dispatching rules, which are reactive in nature. However, rule performance varies according to shop utilisation and due date allowance, which often change in dynamic real-world job shops. Therefore, this paper explores systems that select dispatching rules as conditions change over time, namely periodic and real-time dispatching rule selection systems, which are based on supervised learning and reinforcement learning algorithms, respectively. These types of systems have been proposed in the past but have been further improved in this work by carefully selecting the most relevant state features and dispatching rules. Moreover, by testing both approaches on the same instances, it was possible to compare them and determine the most advantageous one. After the tests, which included a wide array of job shop instances, both periodic and real-time systems outperformed state-of-the-art dispatching rules by over 10% tardiness-wise. Nonetheless, the periodic rule selection approach was more robust across all tests than the real-time approach. These results demonstrate that there is a real incentive for managers to adopt dispatching rule selection systems.

Abstract
This paper presents an optimisation approach based on mixed-integer programming for tactical planning decisions within fresh fruit processing industries. It applies to fruits such as oranges, tomatoes, guavas and others, where diluted fruit juice needs to be concentrated in evaporators to produce semi-finished or finished products. It considers agricultural and industrial activities, integrating them to address complex and interconnected decisions. Agricultural tasks include planting, harvesting, and transporting fruits from fields to processing plants, while industrial activities involve the production, inventory, and transportation of semi-finished and final products. This approach accommodates multiple agricultural regions, fruit varieties, processing plants, and products, operating on a weekly basis within a one-year planning horizon. It offers a detailed solution for harvesting, the fruit juice concentration process, inventory management for the products produced, and transportation of raw materials and products among processing plants. Production of semi-finished products is modelled using the Proportional Lot-Sizing and Scheduling Problem and the production of finished products is modelled adopting a blending lot-sizing problem. The results were validated through computational experiments using a dataset from a company that processes tomatoes and guavas. Scenario analyses were conducted to evaluate the solution's consistency and real-world applicability. The findings indicate that the approach can support decision making in practice, highlighting its potential as a valuable managerial, analytical, and optimisation tool for some agri-food industries. © 2025 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.

Abstract
Hydropower currently generates more than all other renewable energies combined. Considering the challenges of climate change and the transition to green energy, it is expected to remain the world's largest source of renewable electricity generation. This paper proposes a tool for performance evaluation and benchmarking of hydropower generation to inform dispatching. Through them, strengths and weaknesses of asset operations can be set, identifying areas with the best performance, gathering insights from their strategies and best practices, and comprehending factors that lead to variations in performance levels. The results allow for optimising energy resource use by indicating the dispatching rules with maximum power production and minimum wearand-tear impact. This framework allows the formulation of practical guidelines for dispatching policies. The proposed methodology is applied to analyse two real-world case studies: the Vogelgr & uuml;n run of river hydropower plant (France) and the Frades 2 pump-storage powerplant (Portugal).

Abstract
Maintaining process quality is one of the biggest challenges manufacturing industries face, as production processes have become increasingly complex and difficult to monitor effectively in today's manufacturing contexts. Reliance on skilled operators can result in suboptimal solutions, impacting process quality. In doing so, the importance of quality monitoring and diagnosis methods cannot be undermined. Existing approaches have limitations, including assumptions, prior knowledge requirements, and unsuitability for certain data types. To address these challenges, we present a novel unsupervised monitoring and detection methodology to monitor and evaluate the evolution of a quality characteristic's degradation. To measure the degradation we created a condition index that effectively captures the quality characteristic's mean and scale shifts from the company's specification levels. No prior knowledge or data assumptions are required, making it highly flexible and adaptable. By transforming the unsupervised problem into a supervised one and utilising historical production data, we employ logistic regression to predict the quality characteristic's conditions and diagnose poor condition moments by taking advantage of the model's interpretability. We demonstrate the methodology's application in a glass container production process, specifically monitoring multiple defective rates. Nonetheless, our approach is versatile and can be applied to any quality characteristic. The ultimate goal is to provide decision-makers and operators with a comprehensive view of the production process, enabling better-informed decisions and overall product quality improvement.