Abstract
A multivariate outlier detection method for interval data is proposed that makes use of a parametric approach to model the interval data. The trimmed maximum likelihood principle is adapted in order to robustly estimate the model parameters. A simulation study demonstrates the usefulness of the robust estimates for outlier detection, and new diagnostic plots allow gaining deeper insight into the structure of real world interval data.

Abstract
This paper addresses new Mixed-model Assembly Line Balancing Problems (MALBP) in a real industrial context, the stitching systems of a footwear company. The work is part of large ongoing projects with this industry, and the main purposes are minimising the number of required workstations and smoothing the operators' workload. The company has invested in new flexible automated assembly systems, which accommodate dozens of workstations and many moving boxes. Footwear components are inside boxes (with various quantities) which can move from the warehouses to a convenient workstation or between any workstations (in any order). This is a significant and distinct feature of the MALBP, together with the fact that the assignment of different skilled operators and machines is achieved simultaneously. An optimisation model is developed, in part to facilitate the understanding of the situation and to solve small-size instances. Due to the complexity of the problems, we had to devise an approximate method, based on the Variable Neighbourhood Descent (VND) metaheuristic and integrating an adaptation of the Ranked Positional Weighted (RPW) method. The adapted RPW method is used to create initial feasible solutions, while preassigning special operators and machines. After choosing good initial solutions, VND is applied to improve their quality. The new contributed method, named as RPW-VNDbal, is tested with medium and large instances, in two distinct stitching systems. A Lower Bound of the objective function and Simulation contribute to evaluate the solutions and their practicability. The results implemented by the project team, show that the RPW-VNDbal method is fast enough and offers better solutions than those implemented by the experienced operation managers of the company.

Abstract

Abstract
Cannabis is the most used illicit substance in Portugal and is the one, which presents, on average, an earlier onset of consumption. Craving is considered one of the fundamental factors linked to the cannabis use, associating anxiety, stress and depression. The sample consisted of 143 individuals from the northern area of the country and the data was gathered through two online questionnaires (DABS-21 and the MCQ-SF). The obtained results allow us to conclude that emotionality, purposefulness and compulsivity are positively associated with the level of craving and, combined, as well to stress and depression. The number of years of consumption presents a significant effect in anxiety, with a medium effect size.

Abstract
This paper investigates ac-dc-ac multilevel topologies for three-phase four-wire applications obtained from an interconnection of three-leg voltage-source converters. The topologies can be applied in line voltage regulators, universal active power filters, and uninterrupted power supplies. Operating principles, a pulsewidth modulation (PWM) technique, and an overall control strategy are developed. The PWM and control strategies are suitable to regulate the dc-link voltages and to reduce the harmonic distortion, switching stress, and power losses. Simulation results are used to compare the proposed configurations with a conventional solution in terms of harmonic distortion, semiconductor losses, and others characteristics. Experimental results are presented to validate the theoretical considerations and were obtained by using insulated-gate bipolar transistors with dedicated drives and a digital signal processor with appropriated plug-in boards and sensors.

Abstract