Decision & Information Sciences for Production Systems

Clustering algorithms have emerged as an alternative powerful meta-learning tool to accurately analyze the massive volume of data generated by modern applications. In particular, their main goal is to categorize data into clusters such that objects are grouped in the same cluster when they are similar according to specific metrics. There is a vast body of knowledge in the area of clustering and there has been attempts to analyze and categorize them for a larger number of applications. However, one of the major issues in using clustering algorithms for big data that causes confusion amongst practitioners is the lack of consensus in the definition of their properties as well as a lack of formal categorization. With the intention of alleviating these problems, this paper introduces concepts and algorithms related to clustering, a concise survey of existing (clustering) algorithms as well as providing a comparison, both from a theoretical and an empirical perspective. From a theoretical perspective, we developed a categorizing framework based on the main properties pointed out in previous studies. Empirically, we conducted extensive experiments where we compared the most representative algorithm from each of the categories using a large number of real (big) data sets. The effectiveness of the candidate clustering algorithms is measured through a number of internal and external validity metrics, stability, runtime, and scalability tests. In addition, we highlighted the set of clustering algorithms that are the best performing for big data.

Abstract This work aims to review literature related to the latest cyber-physical systems (CPS) for manufacturing in the revolutionary Industry 4.0 for a comprehensive understanding of the challenges, approaches, and used techniques in this domain. Different published studies on CPS for manufacturing in Industry 4.0 paradigms through 2010 to 2019 were searched and summarized. We, then, analyzed the studies at a different granularity level inspecting the title, abstract, and full text to include in the prospective study list. Out of 626 primarily extracted relevant articles, we scrutinized 78 articles as the prospective studies on CPS for manufacturing in Industry 4.0. First, we analyzed the articles’ context to identify the major components along with their associated fine-grained constituents of Industry 4.0. Then, we reviewed different studies through a number of synthesized matrices to narrate the challenges, approaches, and used techniques as the key-enablers of the CPS for manufacturing in Industry 4.0. Although the key technologies of Industry 4.0 are the CPS, Internet of Things (IoT), and Internet of Services (IoS), the human component (HC), cyber component (CC), physical component (PC), and their HC-CC, CC-PC, and HC-PC interfaces need to be standardized to achieve the success of Industry 4.0.

Abstract Increasingly, humanitarian organizations have opened regional warehouses and pre‐positioned resources locally. Choosing appropriate locations is not easy and frequently based on opportunities rather than rational decisions. Dedicated decision‐support systems could help humanitarian practitioners design their supply networks. Academic literature suggests the use of commercial sector models but rarely considers the constraints and specific context of humanitarian operations, such as obtaining accurate data, high uncertainties, limited budgets and increasing pressure on cost efficiency. We propose a tooled methodology to properly support humanitarian decision makers in the design of their supply chains. Our contribution is based on the definition of aggregate scenarios to reliably forecast demand using past disaster data and future trends. Demand for relief items based on these scenarios is then fed to a mixed‐integer linear programming model in order to improve current supply networks. The specifications of this model have been defined in close collaboration with humanitarian workers. The model allows analysis of the impact of alternative sourcing strategies and service level requirements on operational efficiency. It provides clear and actionable recommendations for a given context, bridging the gap between academics and humanitarian logisticians. The methodology was developed to be useful to a broad range of humanitarian organizations, and a specific application to the supply chain design of the International Federation of Red Cross and Red Crescent Societies is discussed in detail.

Sentiment analysis is an automatic way to determine that whether opinions of people about a subject are favorable or unfavorable. One of the most important sub tasks in sentiment analysis is to determine the sequence of words affected by negation. Most of the existing sentiment analysis systems used traditional methods based on static window and punctuation marks to determine the scope of negation. However, these methods do not offer satisfactory performance due to variability in the negation scope, inability to deal with linguistic features and improper word sense disambiguation. In this paper, we investigate the problem of identifying the scope of negation while determining the polarity of a sentence. We propose a negation handling method based on linguistic features which determine the effect of different types of negation. Experiment results show that the proposed method improves the accuracy of both negation scope identification and overall sentiment analysis.

By connecting devices, people, vehicles, and infrastructures everywhere in a city, governments and their partners can improve community well-being and other economic and financial aspects (e.g., cost and energy savings). Nonetheless, smart cities are complex ecosystems that comprise many different stakeholders (network operators, managed service providers, logistic centers, and so on), who must work together to provide the best services and unlock the commercial potential of the so-called Internet of Things (IoT). This is one of the major challenges that faces today's smart city movement, and the emerging “API economy.” Indeed, while new smart connected objects hit the market every day, they mostly feed “vertical silos” (e.g., vertical apps, siloed apps, and so on) that are closed to the rest of the IoT, thus hampering developers to produce new added value across multiple platforms and/or application domains. Within this context, the contribution of this paper is twofold: (1) present the strategic vision and ambition of the EU to overcome this critical vertical silos' issue and (2) introduce the first building blocks underlying an open IoT ecosystem developed as part of an EU (Horizon 2020) Project and a joint project initiative (IoT-EPI). The practicability of this ecosystem, along with a performance analysis, is carried out considering a proof-of-concept for enhanced sporting event management in the context of the forthcoming FIFA World Cup 2022 in Qatar.

Several recent reviews summarize common missing value analysis methods. However, none of them provide a systematic and in-depth summary of the analytical challenges and solutions for dealing with missing values. For the purpose of guiding the handling of missing values, this review aims to consolidate current developments in novel missing-value research methodologies. In particular, we comprehensively investigated cutting-edge missing value solutions and methodically studied the main challenges associated with missing values analysis (missing mechanisms, missing patterns, and missing rates). Furthermore, we reviewed 63 publications that compare different strategies for deleting and imputing missing values. Then we investigated data characteristics, highlighted three main problems when analyzing missing values, and analyzed the performance of missing value solutions in these studied papers. Moreover, we conducted comprehensive experiments on 9 public datasets using typical missing value processing methods and provided a simple guided decision tree for handling missing values. Finally, we described current Research hotspots and open challenges, which give potential research topics. • Analyzed three major difficulties with missing value analysis. • Provided a comprehensive introduction to deletion and imputation missing approaches. • Reviewed and analyzed numerous studies and provide useful rules for processing missing values. • Conducted experiments and provided a guided decision tree for missing value processing. • Analyzed and summarized the existing research hotspots and open challenges of missing values.

Currently, medical media technologies have become a center of attention due to emerging trends in miniaturized wearable devices from factories to health corner stores everywhere. Due to the power-constrained nature of these portable devices, it is challenging to adopt them during critical medical operations and diagnoses. Maximizing energy efficiency and, hence, extending the battery life is vital. In addition, conventional approaches with constant transmission power are inappropriate option for green and smart healthcare. Thus, this paper first proposes a transmission power control (TPC)-based energy-efficient algorithm (EEA) for when a subject is in different postures, i.e., standing, walking, and running, in wireless body sensor networks. Second, a hardware platform was developed on the Intel Galileo board to test and compare the proposed EEA and conventional adaptive TPC (ATPC) in terms of energy and channel reliability or packet loss ratio (PLR). Experimental results revealed that the proposed EEA obtained energy savings of 42.5% with an acceptable PLR compared with that of the traditional ATPC method.

OBJECTIVE: This systematic review sought to establish a picture of length of stay (LOS) prediction methods based on available hospital data and study protocols designed to measure their performance. MATERIALS AND METHODS: An English literature search was done relative to hospital LOS prediction from 1972 to September 2019 according to the PRISMA guidelines. Articles were retrieved from PubMed, ScienceDirect, and arXiv databases. Information were extracted from the included papers according to a standardized assessment of population setting and study sample, data sources and input variables, LOS prediction methods, validation study design, and performance evaluation metrics. RESULTS: Among 74 selected articles, 98.6% (73/74) used patients' data to predict LOS; 27.0% (20/74) used temporal data; and 21.6% (16/74) used the data about hospitals. Overall, regressions were the most popular prediction methods (64.9%, 48/74), followed by machine learning (20.3%, 15/74) and deep learning (17.6%, 13/74). Regarding validation design, 35.1% (26/74) did not use a test set, whereas 47.3% (35/74) used a separate test set, and 17.6% (13/74) used cross-validation. The most used performance metrics were R2 (47.3%, 35/74), mean squared (or absolute) error (24.4%, 18/74), and the accuracy (14.9%, 11/74). Over the last decade, machine learning and deep learning methods became more popular (P=0.016), and test sets and cross-validation got more and more used (P=0.014). CONCLUSIONS: Methods to predict LOS are more and more elaborate and the assessment of their validity is increasingly rigorous. Reducing heterogeneity in how these methods are used and reported is key to transparency on their performance.

Purpose Digital tools have been used to document cultural heritage with high-quality imaging and metadata. However, some of the historical assets are totally or partially unlabeled and some are physically damaged, which decreases their attractiveness and induces loss of value. This paper introduces a new framework that aims at tackling the cultural data enrichment challenge using machine learning. Design/methodology/approach This framework focuses on the automatic annotation and metadata completion through new deep learning classification and annotation methods. It also addresses issues related to physically damaged heritage objects through a new image reconstruction approach based on supervised and unsupervised learning. Findings The authors evaluate approaches on a data set of cultural objects collected from various cultural institutions around the world. For annotation and classification part of this study, the authors proposed and implemented a hierarchical multimodal classifier that improves the quality of annotation and increases the accuracy of the model, thanks to the introduction of multitask multimodal learning. Regarding cultural data visual reconstruction, the proposed clustering-based method, which combines supervised and unsupervised learning is found to yield better quality completion than existing inpainting frameworks. Originality/value This research work is original in sense that it proposes new approaches for the cultural data enrichment, and to the authors’ knowledge, none of the existing enrichment approaches focus on providing an integrated framework based on machine learning to solve current challenges in cultural heritage. These challenges, which are identified by the authors are related to metadata annotation and visual reconstruction.

In real cases, missing values tend to contain meaningful information that should be acquired or should be analyzed before the incomplete dataset is used for machine learning tasks. In this work, two algorithms named jointly fuzzy C-Means and vaguely quantified nearest neighbor (VQNN) imputation (JFCM-VQNNI) and jointly fuzzy C-Means and fitted VQNN imputation (JFCM-FVQNNI) have been proposed by considering clustering conception and sufficient extraction of uncertain information. In the proposed JFCM-VQNNI and JFCM-FVQNNI algorithm, the missing value is regarded as a decision feature, and then, the prediction is generated for the objects that contain at least one missing value. Specially, as for JFCM-VQNNI algorithm, indistinguishable matrixes, tolerance relations, and fuzzy membership relations are adopted to identify the potential closest filled values based on corresponding similar objects and related clusters. On the basis of JFCM-VQNNI algorithm, JFCM-FVQNNI algorithm synthetic analyzes the fuzzy membership of the dependent features for instances with each cluster. In order to fill the missing values more accurately, JFCM-FVQNNI algorithm performs fuzzy decision membership adjustment in each object with respect to the related clusters by considering highly relevant decision attributes. The experiments have been carried out on five datasets. Based on the analysis of root-mean-square error, mean absolute error, comparison of imputation values with actual values, and classification accuracy results analysis, we can draw the conclusion that the proposed JFCM-FVQNNI and JFCM-VQNNI algorithms yields sufficient and reasonable imputation performance results by comparing with fuzzy C-Means parameter-based imputation algorithm and fuzzy C-Means rough parameter-based imputation algorithm.

International audience

The automated and timely conversion of cybersecurity information from unstructured online sources, such as blogs and articles to more formal representations has become a necessity for many applications in the domain nowadays. Named Entity Recognition (NER) is one of the early phases towards this goal. It involves the detection of the relevant domain entities, such as product, version, attack name, etc. in technical documents. Although generally considered a simple task in the information extraction field, it is quite challenging in some domains like cybersecurity because of the complex structure of its entities. The state of the art methods require time-consuming and labor intensive feature engineering that describes the properties of the entities, their context, domain knowledge, and linguistic characteristics. The model demonstrated in this paper is domain independent and does not rely on any features specific to the entities in the cybersecurity domain, hence does not require expert knowledge to perform feature engineering. The method used relies on a type of recurrent neural networks called Long Short-Term Memory (LSTM) and the Conditional Random Fields (CRFs) method. The results we obtained showed that this method outperforms the state of the art methods given an annotated corpus of a decent size.

Fog computing has become the revolutionary paradigm and one of the intelligent services of the 5th Generation (5G) emerging network, while Internet of Things (IoT) lies under its main umbrella. Enhancing and optimizing the quality of service (QoS) in Fog computing networks is one of the critical challenges of the present. In the meantime, strong links between the Fog, IoT devices and the supporting back-end servers is done through large scale cloud data centers and with the linear exponential trend of IoT devices and voluminous generated data. Fog computing is one of the vital and potential solutions for IoT in close connection with things and end users with less latency but due to high computational complexity, less storage capacity and more power drain in the cloud it is inappropriate choice. So, to remedy this issue, we propose transmission power control (TPC) based QoS optimization algorithm named (QoS-TPC) in the Fog computing. Besides, we propose the Fog-IoT-TPC-QoS architecture and establish the connection between TPC and Fog computing by considering static and dynamic conditions of wireless channel. Experimental results examine that proposed QoS-TPC optimizes the QoS in terms of maximum throughput, less delay, less jitter and minimum energy drain as compared to the conventional that is, ATPC, SKims and constant TPC methods.

The good running of a Home Health Care Structure implies the resolution of several complex problems. In this literature review, we give an overview of current work dealing with planning activities in Home Health Care Structures, and more particularly routing and scheduling of caregivers. Since the patient is the heart of the decision-making system, the reviewed works often consider non-standard criteria and constraints, compared with other routing problems. We focus more particularly on these non-standard and practical aspects of the considered problems.

In recent years, implementing coordination mechanisms in decentralised supply chains to reduce the well-known negative effects of decentralisation, such as the ‘bullwhip effect’, has become a considerable challenge. Furthermore, with the dramatic developments in information and communication technologies, real-time information sharing has become increasingly easier to implement. In this work, we study a mono-product divergent supply chain composed of a supplier, a warehouse, retailers and customers in the context of decentralised and centralised decisions. The main objective of this study is to compare a decentralised supply chain combined with different scenarios of simultaneous upstream and downstream information sharing vs. a centralised supply chain. A mathematical model is developed to compare the logistics costs in the two decision contexts. The experimental results clearly show that the simultaneous sharing of customer demand and supplier-warehouse lead time information in a decentralised supply chain yields nearly equivalent logistics costs as the centralised supply chain context. However, the main beneficiary of the sharing is the warehouse, which receives approximately two-thirds of the benefit. Thus, incentives and revenue sharing contracts should be implemented to motivate and balance the benefits between supply chain partners.

International audience

We study self regulation through pricing for Vehicle Sharing Systems (VSS). Without regulation VSS have poor performances. We study possible improvements of the VSS using (pricing as) incentive. We take as base model a Markovian formulation of a closed queuing network with finite buffer and time dependent service time. This model is unfortunately intractable for the size of instances we want to tackle. We present a fluid approximation, constructed by replacing stochastic demands by a continuous deterministic flow (keeping the demand rate). It gives a deterministic and continuous dynamics and evolves as a continuous process. The fluid model can be thought of as the limit of a scaled Markovian model, in which the fleet, the station capacities and the demands are scaled by a common factor. A reusable benchmark and an experimental protocol is created for the general Vehicle Sharing System optimization problem. We discuss the convergence of the scaled model to the fluid limit. We conduce some experimental studies.

International audience

As the main supplier of the workforce to the industry, higher education is increasingly criticized for not being abreast with the digital revolution and being disconnected from the industry. Competency-based education was developed to address this issue and bridge the gap between what the university is producing and the requirements of the industry. Hence, tools need to be developed that assists in the analysis process. This paper focuses on proposing a system that models the competencies required by occupations in the industry and higher education curricula and assists in matching profiles from the two domains. The different concepts in the domain are modeled as a semantic web ontology, and an inference engine performs the profile matching. In addition to the profile matching, the system calculates a score for the matching degree using the analytic hierarchy process (AHP) method.

Cultural heritage takes an important part in defining the identity and the history of a civilization or a nation. Valuing and preserving this heritage is thus a top priority for governments and heritage institutions. Through this paper, we present an image completion (inpainting) approach adapted for the curation and the completion of damaged artwork. Our approach uses a set of machine learning techniques such as Generative Adversarial Networks which are among the most powerful generative models that can be trained to generate realistic data samples. As we are focusing mostly on visual cultural heritage, the pipeline of our framework has many optimizations such as the use of clustering to optimize the training of the generative part to ensure a better performance across a variety of cultural data categories. The experimental results of our framework are promising and were validated on a dataset of paintings.