Abstract
Trip duration is an important metric for the management of taxi companies, as it affects operational efficiency, driver satisfaction and, above all, customer satisfaction. In particular, the ability to predict trip duration in advance can be very useful for allocating taxis to stands and finding the best route for trips. A data mining approach can be used to generate models for trip time prediction. In fact, given the amount of data available, different models can be generated for different taxis. Given the difference between the data collected by different taxis, the best model for each one can be obtained with different algorithms and/or parameter settings. However, finding the configuration that generates the best model for each taxi is computationally very expensive. In this paper, we propose the use of metalearning to address the problem of selecting the algorithm that generates the model with the most accurate predictions for each taxi. The approach is tested on data collected in the Drive-In project. Our results show that metalearning can help to select the algorithm with the best accuracy.

Abstract
The present invention is directed to manufacturing, wherein each component of the outsourced process and system can be remotely controlled in decentralized manufacturing process, integrating resources and stakeholders in a global chain, that utilizes ubiquitous computing systems virtual and networked enterprises concepts, for anywhere-anytime control. Experiment involved 68 subjects, Serbian students that have used the interface for remote collaborative control to control CNC machine located in Portugal. Usability evaluation measures of the distributed remote user controlled manufacturing system, such as percentage of task completed, its accuracy and time to complete task, are measured on two types of "client" user interface ("Wall" and "Window"), in two modes of presentation (desktop and video beam), working individually or in small collaborative group consisted of two persons, results were statistically tested. Percentage of task completed is equal for both types of interfaces, while accuracy of task is significantly better for "Wall" interface. Time of task execution is longer for individuals when desktop is used comparing to video beam. There exist significant differences when work is conducted in groups via video beam and time is longer when desktop is used. Also, work time is longer with "Wall" interface. When working in group, work time is shorter than the time when working individually, whereby the working time on desktop is longer than on the video beam. Also, the time is longer when working individually using video beam, using "Wall" interface on desktop, and "Window" interface on video beam. These results show that group work consumes less time in most working options, giving the best results when working in collaborative small group on "Wall" interface via video beam. (C) 2015 The Authors. Published by Elsevier B.V.

Abstract
Medical devices are becoming more interconnected and complex, and are increasingly supported by fragmented organizational systems, e.g. through different processes, committees, supporting staff and training regimes. Distributed Cognition has been proposed as a framework for understanding the design and use of medical devices. However, it is not clear that it has the analytic apparatus to support the investigation of such complexities. This paper proposes a framework that introduces concentric layers to DiCoT, a method that facilitates the application of Distributed Cognition theory. We use this to explore how an inpatient blood glucose meter is coupled with its context. The analysis is based on an observational study of clinicians using a newly introduced glucometer on an oncology ward over approximately 150. h (11. days and 4 nights). Using the framework we describe the basic mechanics of the system, incremental design considerations, and larger design considerations. The DiCoT concentric layers (DiCoT-CL) framework shows promise for analyzing the design and use of medical devices, and how they are coupled with their context. © 2014 Elsevier Inc.

Abstract
This paper presents the analysis of a distribution system subject to reconfiguration with high wind penetration over a period of 24 hours. In order to meet this objective, the reconfiguration problem is solved through mixed integer linear programming considering the stochasticity of the variables, where the balance between load and generation has to be satisfied at the lowest cost in each period. The model considers a wind profile that follows an Auto-Regressive (AR) pattern to take into account uncertainty in wind generation, bus loads for a typical demand profile, and non-renewable distributed generation, all implemented for a 33-bus case study representing a weakly meshed distribution network. Test results from the case study show the effects on voltages, currents and substation behaviors as well as the power loss and the overall cost of the system.

Abstract

Abstract