Global supply chain management: A reinforcement learning approach

In recent years, researchers and practitioners alike have devoted a great deal of attention to supply chain management (SCM). The main focus of SCM is the need to integrate operations along the supply chain as part of an overall logistic support function. At the same time, the need for globalization requires that the solution of SCM problems be performed in an international context as part of what we refer to as Global Supply Chain Management (GSCM). This paper proposes an approach to study GSCM problems using an artificial intelligence framework called reinforcement learning (RL). The RL framework allows the management of global supply chains under an integration perspective. The RL approach has remarkable similarities to that of an autonomous agent network (AAN); a similarity that we shall discuss. The RL approach is applied to a case example, namely a networked production system that spans several geographic areas and logistics stages. We discuss the results and provide guidelines and implications for practical applications.     

The scientific economy of attention: A novel approach to the collective rationality of science

Science is the core sector of present-day knowledge production. Yet, the mechanisms of science as an industry are poorly understood. The economic theory of science is still in its infancy, and philosophy of science has only sparsely addressed the issue of economic rationality. Research, however, is costly. Inefficient use of resources consumed by the scientific industry is as detrimental to the collective advancement of knowledge as are deficiencies in method. Economic inefficiency encompasses methodological inadequacy. Methods are inadequate if they tend to misallocate time and effort. If one omits the question of how inputs are transformed into outputs in self-organised knowledge production, this means neglecting an essential aspect of the collective rationality of science. A self-organised tendency towards efficiency comes to the fore as soon as science is described as an economy in which researchers invest their own attention in order to obtain the attention of others. Viewed like this, scientific communication appears to be a market where information is exchanged for attention. Scientific information is measured in terms of the attention it earns. Since scientists demand scientific information as a means of production, the attention that a theory attracts is a measure of its value as a capital good. On the other hand, the attention a scientist earns is capitalised into the asset called reputation. Elaborating the ideas introduced in Franck (1998) and (1999), the paper describes science as a highly developed market economy. Science conceived as capital market covers the specific conditions under which scientists, while maximising their reputation, optimise output in the eyes of those competent to judge. Attention is not just any resource. It is the resource whose efficient use is called intelligence. Science, as an industry transforming attention into cognitive output, is bound to miss the hallmark of rationality if it does not pass a test of collective intelligence. The paper closes with considering the prospective outcome of such a test. This revised version was published online in June 2006 with corrections to the Cover Date.     

Design of a biochemical circuit motif for learning linear functions

Learning and adaptive behaviour are fundamental biological processes. A key goal in the field of bioengineering is to develop biochemical circuit architectures with the ability to adapt to dynamic chemical environments. Here, we present a novel design for a biomolecular circuit capable of supervised learning of linear functions, using a model based on chemical reactions catalysed by DNAzymes. To achieve this, we propose a novel mechanism of maintaining and modifying internal state in biochemical systems, thereby advancing the state of the art in biomolecular circuit architecture. We use simulations to demonstrate that the circuit is capable of learning behaviour and assess its asymptotic learning performance, scalability and robustness to noise. Such circuits show great potential for building autonomous in vivo nanomedical devices. While such a biochemical system can tell us a great deal about the fundamentals of learning in living systems and may have broad applications in biomedicine (e.g. autonomous and adaptive drugs), it also offers some intriguing challenges and surprising behaviours from a machine learning perspective.     

A fuzzy group decision making approach to new product concept screening at the fuzzy front end

Efficient new product concept screening plays an important role in new product development (NPD). Due to the inherent complexity, incompleteness and ambiguity in NPD, fuzzy approaches have been widely developed and used for new product concept screening. Because of the group behaviours (diversity of gate-team members and senior management’s preference), the new product concept screening is typical group decision-making problem. Moreover, it is usually more efficient to conduct the screening evaluation in a gate-team and make a screening decision outside the gate-team by senior management. Therefore, it is necessary to provide a result with a high degree of flexibility and credibility so as to better support the senior management to make a screening decision in NPD. Existing fuzzy approaches neglect these two issues. Towards this end, a three-stage fuzzy group decision-making approach is proposed to assist the senior management perform new product concept screening. A case study is borrowed from the literature to illustrate our proposed approach. Our approach is compared with two existing techniques from theoretical and empirical perspectives. The results show that the proposed approach can capture the group behaviours and present a relatively realistic and informative result which gives the senior management a high degree of flexibility and credibility in new product concept screening.     

Deep Learning Control for Autonomous Robot

Several applications of machine learning and artificial intelligence, have acquired importance and come to the fore as a result of recent advances and improvements in these approaches. Autonomous cars are one such application. This is expected to have a significant and revolutionary influence on society. Integration with smart cities, new infrastructure and urban planning with sophisticated cyber-security are some of the current ramifications of self-driving automobiles. The autonomous automobile, often known as self-driving systems or driverless vehicles, is a vehicle that can perceive its surroundings and navigate predetermined routes without human involvement. Cars are on the verge of evolving into autonomous robots, thanks to significant breakthroughs in artificial intelligence and related technologies, and this will have a wide range of socio-economic implications. However, in order for these automobiles to become a reality, they must be endowed with the perception and cognition necessary to deal with high-pressure real-life events and make proper judgments and take appropriate action. The majority of self-driving car technologies are based on computer systems that automate vehicle control parts. From forward-collision warning and antilock brakes to lane-keeping and adaptive drive control, to fully automated driving, these technological components have a wide range of capabilities. A self-driving car combines a wide range of sensors, actuators, and cameras. Recent researches on computer vision and deep learning are used to control autonomous driving systems. For self-driving automobiles, lane-keeping is crucial. This study presents a deep learning approach to obtain the proper steering angle to maintain the robot in the lane. We propose an advanced control for a self-driving robot by using two controllers simultaneously. Convolutional neural networks (CNNs) are employed, to predict the car’ and a proportional-integral-derivative (PID) controller is designed for speed and steering control. This study uses a Raspberry PI based camera to control the robot car.     

A discrete-time queueing network approach to performance evaluation of autonomous vehicle storage and retrieval systems

In this paper, we present a method for performance evaluation of autonomous vehicle storage and retrieval systems (AVS/RSs) with tier-captive single-aisle vehicles. A discrete-time open queueing network approach is applied. The data obtained from the evaluation of the lift and vehicle movements can be used directly as input for the general discrete service time distributions of the queueing network. Furthermore, the approach allows for the computation of the retrieval transaction time distribution as well as of the distribution of the number of transactions waiting to be stored. Consequently, not only expected values and variances but also quantiles of the performance measures can be obtained. Comparison to discrete-event simulation quantifies approximation errors resulting from the decomposition approach in the discrete-time domain. Moreover, the errors obtained by the discrete-time approach are compared to the errors obtained using a continuous-time open queueing network approach. Finally, it will be outlined how the model can be used for designing AVS/RSs according to given system requirements, such as storage capacity, throughput, height and length of the system as well as the 95% quantile of the retrieval transaction time.     

Living Materials Herald a New Era in Soft Robotics

Living beings have an unsurpassed range of ways to manipulate objects and interact with them. They can make autonomous decisions and can heal themselves. So far, a conventional robot cannot mimic this complexity even remotely. Classical robots are often used to help with lifting and gripping and thus to alleviate the effects of menial tasks. Sensors can render robots responsive, and artificial intelligence aims at enabling autonomous responses. Inanimate soft robots are a step in this direction, but it will only be in combination with living systems that full complexity will be achievable. The field of biohybrid soft robotics provides entirely new concepts to address current challenges, for example the ability to self‐heal, enable a soft touch, or to show situational versatility. Therefore, “living materials” are at the heart of this review. Similarly to biological taxonomy, there is a recent effort for taxonomy of biohybrid soft robotics. Here, an expansion is proposed to take into account not only function and origin of biohybrid soft robotic components, but also the materials. This materials taxonomy key demonstrates visually that materials science will drive the development of the field of soft biohybrid robotics.     

An alliance of humans and machines for machine learning: Hybrid intelligent systems and their design principles

With the growing number of applications of artificial intelligence such as autonomous cars or smart industrial equipment, the inaccuracy of utilized machine learning algorithms could lead to catastrophic outcomes. Human-in-the-loop computing combines human and machine intelligence resulting in a hybrid intelligence of complementary strengths. Whereas machines are unbeatable in logic and computation speed, humans are contributing with their creative and dynamic minds. Hybrid intelligent systems are necessary to achieve high accuracy and reliability of machine learning algorithms. In a design science research project with a Swedish manufacturing company, this paper presents an application of human-in-the-loop computing to make operational processes more efficient. While conceptualizing a Smart Power Distribution for electric industrial equipment, this research presents a set of principles to design machine-learning algorithms for hybrid intelligence. From being AI-ready as an organization to clearly focusing on the customer benefits of a hybrid intelligent system, designers need to build and strengthen the trust in the human-AI relationship to make future applications successful and reliable. With the growing trends of technological advancements and incorporation of artificial intelligence in more and more applications, the alliance of humans and machines have become even more crucial.     

Scenario-based approach for flexible resource loading under uncertainty

Order acceptance decisions in manufacture-to-order environments are often made based on incomplete or uncertain information. To quote reliable due dates in order processing, manage resource capacity adequately and take into account uncertainty, the paper presents and analyses models and tools for more robust resource loading. We refer to the problem as flexible resource loading under uncertainty. We propose a scenario-based solution approach that can deal with a wide range of uncertainty types. The approach is based on an MILP to find a plan with minimum expected costs over all relevant scenarios. To solve this MILP, we propose an exact branch-and-price algorithm. Further, we propose a much faster improvement heuristic based on an LP (linear programming) approximation. A disadvantage of the scenario-based MILP, is that a large number of scenarios may make the model intractable. We therefore propose an approximate approach that uses the aforementioned solution techniques and only a sample of all scenarios. Computational experiments show that, especially for instances with sufficient slack, solutions obtained with deterministic techniques that only use the expected scenario can be significantly improved with respect to their expected costs (i.e. robustness). We also show that for large instances, our heuristics outperform the exact approach given a maximum computation time as a stopping criterion. Moreover, it turns out that using a small sample of selected scenarios generally yields better results than using all scenarios.     

Formation of autonomous agent networks for manufacturing systems

Designing flexible manufacturing systems that can cope well with the dynamic environment has been an important goal. The objective of this paper is to describe a modelling approach developed to design a manufacturing system as a society of autonomous agents called autonomous agent network (AAN). Within the AAN, autonomous agents are loosely coupled. System's tasks are accomplished by the autonomous agents collaboratively through the communication and information exchange definitions protocols. The approach in this paper consists of autonomous agent formation and protocol formation. Reducing the degree of uncertainty, reducing the impact of uncertainty, and keeping the desired level of productivity are the main motivations for forming autonomous agents for manufacturing. By using two basic communication acts, 'request' and 'tell', five basic protocols are formed. The five basic protocols can further form specific taskapplication protocols to define the complex communication among autonomous agents. The methodology is demonstrated with an industrial case study. In addition, the validation of the performance in communication, autonomy and flexibility of AANs are also explained in this paper.     

Lung computed axial tomography image segmentation using possibilistic fuzzy C-means approach for computer aided diagnosis system

Soft computing is an associate rising field that plays a crucial half in the area of engineering and science. One of the most significant applications of soft computing is image segmentation. It focuses on an exploiting tolerance of imprecision and uncertainty. Segmentation supported soft computing remains a difficult task within the medical field. Medical images are habitually used in the segmentation process to extract the meaningful portions and to know and clarify the condition of the particular patient. In this article, we implement an efficient possibilistic fuzzy C-means (PFCM) approach to segment the lung portion in the computed tomography (CT) image and the result shows that it improves the segmentation accuracy upto 98.5012% and results are compared with existing segmenting approaches like fuzzy possibilistic C-means method, fuzzy bitplane method and so forth. Also, the PFCM approach increases the diagnostic accuracy of the computer aided diagnosis system using CT images. The radiologist may utilize this computer aided diagnosis system results as a second opinion of their diagnosed results.     

A hierarchical approach to simulate the packing density of particle mixtures on a computer

In many fields of materials science it is important to know how densely a particle mixture can be packed. The “packing density” is the ratio of the particle volume and the volume of the surrounding container needed for a random close packing of the particles. We present a method for estimating the packing density for spherical particles based on computer simulations only, i.e. without the need for additional experiments. Our method is particularly suited for particle mixtures with an extremely wide range of particle diameters as they occur e.g. in modern concrete mixtures. A single representative sample from such mixtures would be much larger than can be handled on present standard computers. In our hierarchical approach the diameter range is therefore divided into smaller intervals. Samples from these limited diameter intervals are drawn and their packing density is estimated from a simulated packing. The results are used to “fill” the interstices in the sample from the next larger particle interval. To account for the interaction between particles of different sizes we include larger particles into the sample of smaller ones. The larger ones act as part of the boundary during the packing. Thus we obtain more realistic estimates of how dense a fraction of particles can be packed within the whole mixture. The focus of this paper is on the divide-and-conquer approach and on how the simulation results from the fractions can be collected into an overall estimate of the packing density. We do not go into details of the simulation technique for the single packing. We compare our results to some experimental data to show that our method works at least as good as the classical analytical models like CPM without the need for any experiments.     

Damage analysis of laminated composites using a new coupled micro‐meso approach

In this study, the simplicity and strong physical meaning of micromechanics approach and capability of mesomechanics approach for damage analysis of structures with complex loadings are employed to develop a new micro‐meso approach. For this purpose, a new micromechanics model is developed to predict the matrix cracking initiation and evolution in laminated composites. These damage initiation and evolution are replaced with the damage criteria and flow rule in the continuum damage approach, respectively. The results of this procedure are used in the FEM damage analyses of laminated composites to predict constitutive response of layered composites. It is shown that, the predicted stress distribution and strain energy in a lamina unit cell are in good agreement with the finite element results. Furthermore, it is shown that the predicted stress–strain behaviours are in good agreement with the available experimental results for various laminates with different lay‐ups.     

Tomography approach for multi-object adaptive optics

Multi-object adaptive optics (MOAO) is a solution developed to perform a correction by adaptive optics (AO) in a science large field of view. As in many wide-field AO schemes, a tomographic reconstruction of the turbulence volume is required in order to compute the MOAO corrections to be applied in the dedicated directions of the observed very faint targets. The specificity of MOAO is the open-loop control of the deformable mirrors by a number of wavefront sensors (WFSs) that are coupled to bright guide stars in different directions. MOAO calls for new procedures both for the cross registration of all the channels and for the computation of the tomographic reconstructor. We propose a new approach, called "Learn and Apply (L&A)", that allows us to retrieve the tomographic reconstructor using the on-sky wavefront measurements from an MOAO instrument. This method is also used to calibrate the registrations between the off-axis wavefront sensors and the deformable mirrors placed in the science optical paths. We propose a procedure linking the WFSs in the different directions and measuring directly on-sky the required covariance matrices needed for the reconstructor. We present the theoretical expressions of the turbulence spatial covariance of wavefront slopes allowing one to derive any turbulent covariance matrix between two wavefront sensors. Finally, we discuss the convergence issue on the measured covariance matrices, we propose the fitting of the data based on the theoretical slope covariance using a reduced number of turbulence parameters, and we present the computation of a fully modeled reconstructor.     

A versatile approach for the processing of polymer nanocomposites with self-assembled nanofibre templates

The incorporation of nanoparticles into polymers is a design approach that is used in many areas of materials science. The concept is attractive because it enables the creation of materials with new or improved properties by mixing multiple constituents and exploiting synergistic effects. One important technological thrust is the development of structural materials with improved mechanical and thermal characteristics. Equally intriguing is the possibility to design functional materials with unique optical or electronic properties, catalytic activity or selective permeation. The broad technological exploitation of polymer nanocomposites is, however, stifled by the lack of effective methods to control nanoparticle dispersion. We report a simple and versatile process for the formation of homogeneous polymer/nanofibre composites. The approach is based on the formation of a three-dimensional template of well-individualized nanofibres, which is filled with any polymer of choice. We demonstrate that this template approach is broadly applicable and allows for the fabrication of otherwise inaccessible nanocomposites of immiscible components.     

Design of untethered soft material micromachine for life-like locomotion

Living organisms composed of composite materials with complex structures support autonomous and intelligent behaviors, such as motility, perception and response to changes of the environment. By studying the biological structures and their environmental interactions, researchers are now using these natural systems as models for building soft material machines. In this review, we discuss materials and machine engineering principles to achieve life-like locomotion and functionalities in untethered soft micromachines. Through the various mechanochemical or physical mechanisms, we show how molecular motion can be collectively amplified into versatile macroscopic deformation by materials engineering across multiple length scales. In controlled ways, mobile micromachines are made to crawl, roll or jump and adaptive to various terrains, typically inspired by the terrestrial animals while propulsion of swimming micromachines are guided by aquatic organisms. Besides, out-of-equilibrium behaviors of living systems, such as cell cycling, have stimulated the design of autonomous movement. Furthermore, we review the recent efforts on robotic locomotion intelligence to achieve adaptive, functional locomotion and navigation in complex environment. We finally provide a critical perspective for the field of soft micromachines, and highlight the key challenges of different material systems that need to be overcome to realize practical use.     

A hybrid approach for oestrogen receptor status evaluation

Abstract

The aim of this work is to assist pathologists in the evaluation of tumour cells in microscopic breast images where we distinguish three kinds of cells: positive tumour cells for oestrogen receptor, negative tumour cells for oestrogen receptor, and non-tumour cells. This work has proven to be very difficult because of the variability of cells’ size, shape (morphology) and distribution. Conventional methods for segmentation like thresholding and edge detection are unable to resolve this problem. The herein proposed method is a hybrid approach combining segmentation and classification to ensure better results. While the morphological processes are used for artefact elimination and cell segmentation, the classification algorithm is used to automatically classify all existing cells in the image. The paper contains also a comparative study between fuzzy c-means clustering algorithm and neural network-based classification. The proposed approach was applied on several microscopic breast cancer cells images corresponding to eight patients. The experimental results are efficient and the found values are near to those announced by experts. To better interpret these results, we performed a statistical analysis in terms of sensitivity, specificity and accuracy of detected tumour cells. The statistics proved the efficacy of the proposed approach since a percentage exceeding 90% was recorded for sensitivity, specificity and accuracy for the totality of the studied images. When using neural networks, the statistics are slightly above those gathered with fuzzy c-means. We recorded over 97% for sensitivity, specificity and accuracy of detected cells, reaching an error rate below 3%. Furthermore, it should be kept in mind that analysing breast cells images using the proposed approach gives us important information such as number of tumour cells, and number and percentage of positive tumour cells. Moreover, it is so much less time-consuming than experts’ evaluation.     

Is it opened or closed? Colombian science on the move

Using recent original data from three different sources, the article exhibits some strengths and weaknesses of science in Colombia. It shows that research in this country is in a process of growth although recent results of this positive trend are still to be confirmed. Comparing the evolution of science in Colombia with that of Latin America as a whole, describing and explaining its geographical and institutional concentration as well as its thematic distribution, it also reveals the interdependance between science production dynamics and international cooperation programmes. A basic argument is that the development of science in this country, even though it is fragile and erratic, does not lack sound bases. The indicators used suggest indeed an autonomous scientific motion and inspiration which does not contradict the internationalization process of Colombian science but rectifies the picture of an excessively isolated or dependent community that used to be portrayed.     

A Bayesian approach to diagnosis and prognosis using built-in test

Accounting for the effects of test uncertainty is a significant problem in test and diagnosis, especially within the context of built-in test. Of interest here, how does one assess the level of uncertainty and then utilize that assessment to improve diagnostics? One approach, based on measurement science, is to treat the probability of a false indication [e.g., built-in-test (BIT) false alarm or missed detection] as the measure of uncertainty. Given the ability to determine such probabilities, a Bayesian approach to diagnosis, and by extension, prognosis suggests itself. In the following, we present a mathematical derivation for false indication and apply it to the specification of Bayesian diagnosis. We draw from measurement science, reliability theory, signal detection theory, and Bayesian decision theory to provide an end-to-end probabilistic treatment of the fault diagnosis and prognosis problem.