Abstract
A new approach to gas leakage detection in high pressure distribution networks is proposed, where two leakage detectors are modelled as a linear parameter varying (LPV) system whose scheduling signals are, respectively, intake and offtake pressures. Running the two detectors simultaneously allows for leakage location. First, the pipeline is identified from operational data, supplied by REN-Gasodutos and using an LPV systems identification algorithm proposed in [1]. Each leakage detector uses two Kalman filters where the fault is viewed as an augmented state. The first filter estimates the flow using a calculated scheduling signal, assuming that there is no leakage. Therefore it works as a reference. The second one uses a measured scheduling signal and the augmented state is compared with the reference value. Whenever there is a significant difference, a leakage is detected. The effectiveness of this method is illustrated with an example where a mixture of real and simulated data is used.

Abstract
Multi-agent systems (MAS) are driving the way to design and engineer control solutions that exhibit flexibility, adaptation and reconfigurability, which are important advantages over traditional centralized systems. The understanding, design and testing of such distributed agent-based approaches, and particularly those exhibiting self-* properties, are usually a hard task. Simulation assumes a crucial role to analyse the behaviour of MAS solutions during the design phase and before its deployment into the real operation. Particularly, Agent-Based Modelling (ABM) tools are well suited to simulate MAS systems that exhibit complex phenomena, like emergent behaviour and self-organization. This paper discusses the simulation of agent-based manufacturing systems and introduces the advantages of using ABM tools. The NetLogo platform is used to illustrate the benefits of such tools in the manufacturing world on the specification of a MAS system for a washing machine production line.

Abstract
This paper proposes a heuristic with stochastic neighborhood structures (SNS) to solve two-stage and three-stage two-dimensional guillotine bin packing and cutting stock problems. A solution is represented as a sequence of items which are packed into existing or new stacks, shelves or bins according to previously defined criteria. Moreover, SNS comprises three random neighborhood structures based on modifying the current sequence of items. These are called cut-and-paste, split, and swap blocks and are applied one by one in a fixed order to try to improve the quality of the current solution. Both benchmark instances and real-world instances provided by furniture companies were utilized in the computational tests. Particularly, all benchmark instances are bin packing instances (i.e., the demand of each item type is small), and all real-world instances are classified into bin packing instances and cutting stock instances (i.e., the demand of each item type is large). The computational results obtained by the proposed method are compared with lower bounds and with the solutions obtained by a deterministic Variable Neighborhood Descent (VND) meta-heuristic. The SNS provide solutions within a small percentage of the optimal values, and generally make significant improvements in cutting stock instances and slight to moderate improvements in bin packing instances over the VND approach.

Abstract
Ordinal classification is a form of multiclass classification for which there is an inherent order between the classes, but not a meaningful numeric differerence between them. The performance of such classifiers is usually assessed by measures appropriate for nominal classes or for regression. Unfortunately, these do not account for the true dimension of the error. The goal of this work is to show that existing measures for evaluating ordinal classification models surffer from a number of important shortcomings. For this reason, we propose an alternative measure defined directly in the confusion matrix. An error coefficient appropriate for ordinal data should capture how much the result diverges from the ideal prediction and how "inconsistent" the classifier is in regard to the relative order of the classes. The proposed coefficient results from the observation that the performance yielded by the Misclassification Error Rate coefficient is the benefit of the path along the diagonal of the confusion matrix. We carry out an experimental study which confirms the usefulness of the novel metric.

Abstract
Traditional centralized manufacturing structures were found inadequate to face the challenging requirements of flexibility and re-configurability. Lately, several manufacturing paradigms were introduced to face this challenge, being unified in the objective of providing decentralized control over distributed entities. In spite of their potential benefits, some important questions are far from been answered, namely how control structures are dynamically formed and evolved and how to combine adaptation and optimization. This paper introduces the main principles for re-configurable manufacturing systems that answers to these questions, based on the ADACOR holonic architecture and incorporating mechanisms inspired in other areas of science, notably biology, nature, theory of complexity and artificial life.

Abstract
The current paper proposes a new parametric local color correction technique. Initially, several color transfer functions are computed from the output of the mean shift color segmentation algorithm. Secondly, color influence maps are calculated. Finally, the contribution of every color transfer function is merged using the weights from the color influence maps. The proposed approach is compared with both global and local color correction approaches. Results show that our method outperforms the technique ranked first in a recent performance evaluation on this topic. Moreover, the proposed approach is computed in about one tenth of the time.