Abstract

Abstract
Experimenting is one of the basic tools for design and improvement of products and manufacturing processes. Experiments conducted in industrial conditions should marginally interfere with course of a studied process. These expectations are fulfilled in the best way by so-called passive experiments, in which data is gathered using records obtained from processes conducted in their natural environment and real conditions, without any alteration of set-up parameters. However, it is difficult to control interference originating from variability of studied factors during course of such experiments. That is why use of passive experiments in practice is limited. It is probably caused by distrust regarding credibility of obtained results. Aim of the paper is to compare results obtained from passive and active experiments, conducted in comparable conditions during a process of manufacturing of a selected product and presentation of guidelines for application of passive experiments, developed on the basis of the obtained results.

Abstract
This paper looks as initial results of research into the validity of application of selected priority rules in the development of a levelled production plan. The development of a levelled production plan is the key stage of the authors' own methodology of levelling production to mitigate the adverse impact of variable demand. A levelled production plan permits to maximize effects, defined as being able to deliver diverse products in a timely manner and at the same time reduce stocks and optimize efficient use of manufacturing resources. Application of an appropriate priority rule in the development of a levelled production plan is the key factor determining the effectiveness of the developed methodology. Initial research has been conducted for twenty automatically generated task sets. One hundred manufacturing schedules have been developed in total-five schedules for each set, according to the selected priority rules (shortest task time, longest task time, shortest processing time, longest processing time, first in first out). The schedules have been assessed in terms of meeting the selected key criteria for the objectives of levelled production.

Abstract
This paper presents results of a study of the level of agreement of assessments and decisions made by quality controllers in a sensory inspection. The study subjects included persons with hearing disability. The study was conducted in an automotive manufacturing company. The employees with hearing loss perform product quality inspection which includes, among other things, visual inspection of the goods. The measurement system analysis (MSA) procedure for attributes, with the use of Gwet's AC(1) coefficient, was applied to estimate the level of agreement of decisions made by the raters. The study results show that the quality of work performed by employees with hearing disability is at least as good as that of able-bodied employees. Based on the knowledge possessed by the authors, this approach to the problem is characterized by novelty.

Abstract
Objective: Serious Games is a field of research that has evolved substantially with valuable contributions to many application domains and areas. Patients often consider traditional rehabilitation approaches to be repetitive and boring, making it difficult for them to maintain their ongoing interest and assure the completion of the treatment program. Since the publication of our first taxonomy of Serious Games for Health Rehabilitation (SGHR), many studies have been published with game prototypes in this area. Based on literature review, our goal is to propose an updated taxonomy taking into account the works, updates, and innovations in game criteria that have been researched since our first publication in 2010. In addition, we aim to present the validation mechanism used for the proposed extended taxonomy. Materials and Methods: Based on a literature review in the area and on the analysis of the contributions made by other researchers, we propose an extended taxonomy for SGHR. For validating the taxonomy proposal, a questionnaire was designed to use on a survey among experts in the area. Results: An extended taxonomy for SGHR was proposed. As we have identified that, in general, and besides the mechanisms associated with the adoption of a given taxonomy, there were no reported validation mechanisms for the proposals, we designed a mechanism to validate our proposal. The mechanism uses a questionnaire addressed to a sample of researchers and professionals with experience and expertise in domains of knowledge interrelated with SGHR, such as Computer Graphics, Game Design, Interaction Design, Computer Programming, and Health Rehabilitation. Conclusion: The extended taxonomy proposal for health rehabilitation serious games provides the research community with a tool to fully characterize serious games. The mechanism designed for validating the taxonomy proposal is another contribution of this work.

Abstract
In recent years, the artificial intelligence community has taken big strides in the application of reinforcement learning to games or similar environments using deep learning. From Atari to board games, including motor control or riddle solving, fairly generic deep learning algorithms can now achieve great policies by simply learning to play from experience, and minimal knowledge of the specific domain. However, these algorithms are very demanding in terms of time and hardware in order to achieve the results reported in the literature. So much so, that some algorithms would take years to achieve state-of-the-art performance in commodity hardware. Not only that, but even the learning environments can hinder the speed of the learning process, if they have not been performance optimized. In this paper, we evaluate a complex existing environment, and propose a performance-oriented version, which we call GeoFriends2. We describe the motivation behind the creation of our version, and how it is suitable for both single- and multi-agent reinforcement learning. We then use Asynchronous Deep Learning to create complex policies that can act as baselines for future research on this environment. We also describe a set of techniques that speed up the learning process such that tests can be run with commodity hardware in hours, and not weeks, and using much simpler network architectures. © 2018 IEEE.