Abstract
The integration of IoT infrastructures across production systems, together with the extensive digitalisation of industrial processes, are drastically impacting manufacturing value chains and the business models built on the top of them. By exploiting these capabilities companies are evolving the nature of their businesses shifting value proposition towards models relying on product servitization and share, instead of ownership. In this paper, we describe the semantic data-model developed to support a digital platform fostering the reintroduction in the loop and optimization of unused industrial capacity. Such data-model aims to establish the main propositions of the semantic representation that constitutes the essential nature of the ecosystem to depict their interactions, the flow of resources and exchange of production services. The inference reasoning on the semantic representation of the ecosystem allows to make emerge nontrivial and previously unknown opportunities. This will apply not only to the matching of demand and supply of manufacturing services, but to possible and unpredictable relations. For instance, a particular kind of waste being produced at an ecosystem node can be linked to the requirements for an input material needed in a new product being developed on the platform, or new technologies can be suggested to enhance processes under improvement. The overall architecture and individual ontologies are presented and their usefulness is motivated via the application to use cases.

Abstract
Organizations are nowadays more and more structured in projects in the so-called project-based organizations (PBO). The advantages of PBOs - operational and managerial focus and effectiveness - are counterweight by difficulties in information / knowledge sharing and overall coordination. Project management (PM) applications have been adopted by PBOs with success, mostly at the project level. More innovation-oriented PBOs are keen to experiment and adopt a range of different project supporting applications to optimize several aspects of project management, resulting in “ecosystems” of PM related applications. This paper addresses the problems arising from the implementation of a global project coordination and collaboration application in a research center whose ecosystem of PM applications is extensive. The main challenge has been managing change. An action-research approach was followed to successfully implement the new application while reflecting theoretically on the process and results. The main conclusion is that the requirements elicitation and negotiation are as important as the management of change regarding processes and individual practices. © 2018 The Authors. Published by Elsevier Ltd..

Abstract
Production scheduling in the presence of real-time events is of great importance for the successful implementation of real-world scheduling systems. Most manufacturing systems operate in dynamic environments vulnerable to various stochastic real-time events which continuously forces reconsideration and revision of pre-established schedules. In an uncertain environment, efficient ways to adapt current solutions to unexpected events, are preferable to solutions that soon become obsolete. This reality motivated us to develop a tool that attempts to start filling the gap between scheduling theory and practice. The developed prototype is connected to the MRP software and uses meta heuristics to generate a predictive schedule. Then, whenever disruptions happen, like arrival of new tasks or cancelation of others, the tool starts rescheduling through a dynamic-event module that combines dispatching rules that best fit the performance measures pre-classified by Kano’s model. The proposed tool was tested in an in-depth computational study with dynamic task releases and stochastic execution time. The results demonstrate the effectiveness of the model. © 2020, Springer Nature Switzerland AG.

Abstract
Neural networks have been applied as a machine learning tool in many different areas. Recently, they have gained increased attention with what is now called deep learning. Neural networks algorithms have several parameters that need to be tuned in order to maximize performance. The definition of these parameters can be a difficult, extensive and time consuming task, even for expert users. One approach that has been successfully used for algorithm and parameter selection is metalearning. Metalearning consists in using machine learning algorithm on (meta)data from machine learning experiments to map the characteristics of the data with the performance of the algorithms. In this paper we study how a metalearning approach can be used to obtain a good set of parameters to learn a neural network for a given new dataset. Our results indicate that with metalearning we can successfully learn classifiers from past learning tasks that are able to define appropriate parameters.

Abstract
An holistic approach to decision support systems for intelligent public lighting control, must address both energy efficiency and maintenance. Currently, it is possible to remotely control and adjust luminaries behaviour, which poses new challenges at the maintenance level. The luminary efficiency depends on several efficiency factors, either related to the luminaries or the surrounding conditions. Those factors are hard to measure without understanding the luminary operating boundaries in a real context. For this early stage on preventive maintenance design, we propose an approach based on the combination of two models of the network, wherein each is representing a different but complementary perspective on the classifying of the operating conditions of the luminary as normal or abnormal. The results show that, despite the expected and normal differences, both models have a high degree of concordance in their predictions. © 2020, Springer Nature Switzerland AG.

Abstract
LED-based technology is transforming public lighting networks, favouring smart city innovations. Beyond energy efficiency benefits, LED-based luminaries provide real time stateful data. However, most of the municipalities manage all their luminaries equally, independently of its state or the environmental conditions. Some existing approaches to street lighting management are already considering elementary features such as on-off control and individual dimming based on movement or ambient light. Nevertheless, our vision on public (street) lighting management, goes beyond basic consumption monitoring and dimming control, encompassing: a) adaptive lighting, by considering other potential influence factors such as work temperature of the luminaries or the arrangement of the luminaries on the street; b) Colour tuning, for public safety purposes and; c) emergency behaviour control. This paper addresses the first component (adaptive lighting) influence factors, in the scope of a real scenario in a Portuguese municipality.