Search results

Offline reinforcement learning (RL) acquires effective policies by using prior collected large-scale data, while, in some scenarios, collecting data may be hard because it is time-consuming, expensive and dangerous, i.e. health care, autonomous driving, seeking a more efficient offline RL method. The purpose of the study is to introduce an algorithm, which attempts to sample the high-value transitions in the prioritized buffer, and uniformly sample from the normal experience buffer, improving sample efficiency of offline reinforcement learning, as well as alleviating the “extrapolation error” commonly arising in offline RL.

The authors propose a new structure of experience replay architecture, which consists of double experience replies, a prioritized experience replay and a normal experience replay, supplying samples for policy updates in different training phases. At the first training stage, the authors sample from prioritized experience replay according to the calculated priority of each transitions. At the second training stage, the authors sample from the normal experience replay uniformly. The combination of the two experience replies is initialized by the same offline data set.

The proposed method eliminates out-of-distribution problem in an offline RL regime, and promotes training by leveraging a new efficient experience replay. The authors evaluate their method on D4RL benchmark, and the results reveal that the algorithm can achieve superior performance over the state-of-the-art offline RL algorithm. The ablation study proves that the authors’ experience replay architecture plays an important role in terms of improving final performance, data-efficiency and training stability.

Because of the extra addition of prioritized experience replay, the proposed method increases the computational burden and has the risk of changing data distribution due to the combined sample strategy. Therefore, researchers are encouraged to use the experience replay block effectively and efficiently further.

Offline RL is susceptible to the quality and coverage of pre-collected data, which may be not easy to be collected from specific environment, demanding practitioners to handcraft behavior policy to interact with environment for gathering data.

The proposed approach focuses on the experience replay architecture for offline RL, and empirically demonstrates the superiority of the algorithm on data efficiency and final performance over conservative Q-learning across diverse D4RL tasks. In particular, the authors compare different variants of their experience replay block, and the experiments show that the stages, when to sample from the priority buffer, play an important role in the algorithm. The algorithm is easy to implement and can be combined with any Q-value approximation-based offline RL methods by minor adjustment.

This study aims to deal with the case adaptation problem associated with continuous data by providing a non-zero base solution for knowledge users in solving a given situation.

Firstly, the neighbourhood transformation of the initial case base and the view similarity between the problem and the existing cases will be examined. Multiple cases with perspective similarity or above a predefined threshold will be used as the adaption cases. Secondly, on the decision rule set of the decision space, the deterministic decision model of the corresponding distance between the problem and the set of lower approximate objects under each choice class of the adaptation set is applied to extract the decision rule set of the case condition space. Finally, the solution elements of the problem will be reconstructed using the rule set and the values of the problem's conditional elements.

The findings suggest that the classic knowledge matching approach reveals the user with the most similar knowledge/cases but relatively low satisfaction. This also revealed a non-zero adaptation based on human–computer interaction, which has the difficulties of solid subjectivity and low adaptation efficiency.

In this study the multi-case inductive adaptation of the problem to be solved is carried out by analyzing and extracting the law of the effect of the centralized conditions on the decision-making of the adaptation. The adaption process is more rigorous with less subjective influence better reliability and higher application value. The approach described in this research can directly change the original data set which is more beneficial to enhancing problem-solving accuracy while broadening the application area of the adaptation mechanism.

The examination of the calculation cases confirms the innovation of this study in comparison to the traditional method of matching cases with tacit knowledge extrapolation.

The algorithm models established in this study develop theoretical directions for a multi-case induction adaptation study of tacit knowledge.

This study designs a multi-case induction adaptation scheme by combining NRS and CBR for implicitly knowledgeable exogenous cases. A game-theoretic combinatorial assignment method is applied to calculate the case view and the view similarity based on the threshold screening.

Orientation determination is an essential planning task in additive manufacturing (AM) because it directly affects the part quality, build time, geometric tolerance, fabrication cost, etc. This paper aims to propose a negative feedback decision-making (NFDM) model to realize the personalized design of part orientation in AM process.

NFDM model is constructed by integrating two sub-models: proportional–integral–derivative (PID) negative feedback control model and technique for order preference by similarity to an ideal solution (TOPSIS) decision-making model. With NFDM model, a desired target is first specified by the user. Then, the TOPSIS decision model calculates the “score” for the current part orientation. TOPSIS decision model is modified for ease of control. Finally, the PID controller automatically rotates the part based on the error between the user-specified target and the calculated “score”. Part orientation adjustment is completed when the error is eliminated. Five factors are considered in NFDM model, namely, surface roughness, support structure volume, geometric tolerance, build time and fabrication cost.

The case studies of turbine fan and dragon head indicate that the TOPSIS model can be perfectly integrated with the PID controller. This work extends the proposed model to different AM processes and investigates the feasibility of combining different decision-making models with PID controller and the effects of including various evaluation criteria in the integrated model.

The proposed model innovatively takes the TOPSIS decision-making model and the PID control model as a whole. In this way, the uncontrollable TOPSIS model becomes controllable, so the proposed model can control the TOPSIS model to achieve the user-specified targets.

The purpose of this study is to reduce the gloss of the surface of silk fabrics, by treating the fabrics with tea and matting agent, to imitate the aging and retro effects of silk artefacts.

Silk fabrics were treated with different processing techniques. The aged appearance and surface gloss of the silk fabrics were characterised by sensory analysis, measurement of reflectivity, scanning electron microscopy, measurement of brightness and chroma to identify the influential factors.

The application of matting agent on silk fabrics could reduce the lustre of silk fabrics. Treated with matting agent and tea pigments, silk fabrics could be “aged” to achieve retro effects within a relatively short period of time. A number of other factors and mechanisms that affect the reflectivity of the silk fabrics were also identified.

There is no definite index to evaluate the antique effect of fabrics.

The method developed through this study provided a simple and practical solution to achieving aging and retro effects on silk fabrics.

The method for reducing the lustre of silk fabrics by treating them with matting agent is novel and the finding of the relationships among reflectivity and brightness and chroma is original.

This study aims to reduce the harm of industrial lubricants to consumers. Composite aluminum-based grease (CAG) was prepared, and medical-grade montmorillonite (M-MMT) was used to improve the antiwear performance of the prepared grease.

The influence of the additive (M-MMT) on the tribological performance is mainly investigated using a ball-disc wear tester, and the wear scar surface about the disc was characterized by white light interferometer and electrochemical workstation. Moreover, the cell viability test was used to evaluate the safety of the grease.

The results indicated that for the grease containing 1.5% M-MMT, the average coefficient of friction was reduced by about 46% compared with the CAG, the wear volume of the disc reduced by about 74%. Moreover, CAG and 1.5% M-MMT-containing CAG were proved safety by means of the cell viability test.

The integral properties of CAG can be improved with the medical-grade materials as the additives, while ensuring the safety.

The purpose of this paper is to present a robotic arm that can offer better stiffness than traditional industrial robots for improving the quality of holes in robotic drilling process.

The paper introduces a five-degree of freedom (DOF) robot, which consists of a waist, a big arm, a small arm and a wrist. The robotic wrist is composed of two DOFs of pitching and tilting. A parallelogram frame is used for robotic arms, and the arm is driven by a linear electric cylinder in the diagonal direction. Double screw nuts with preload are used in the ball screw to remove the reverse backlash. In addition, dual-motor drive is applied for each DOF in the waist and the wrist to apply anti-backlash control method for eliminating gear backlash.

The proposed robotic arm has the potential for improving robot stiffness because of its truss structure. The robot can offer better stiffness than industrial robots, which is beneficial to improve the quality of robotic drilling holes.

This paper includes the design of a five-DOF robot for robotic drilling tasks, and the stiffness modeling of the robot is presented and verified by the experiment. The robotic system can be used instead of traditional industrial robots for improving the hole quality to a certain extent.

This study aims to dye silk with natural pigments extract of Coreopsis tinctoria, by treating the fabrics with appropriate mordant under suitable dyeing conditions, to achieve good dyeing depth, fastness and ultraviolet (UV) protection.

Firstly, single factor experiments were used to determine the basic dyeing conditions of Coreopsis tinctoria. The optimal process conditions for direct dyeing were determined through orthogonal experiments. After that, the dyeing with mordant was used. Based on the previously determined optimal process conditions, silk fabrics were dyed with different mordanting methods, with different mordants and mordant dosages. The dyeing results were compared, in terms of the K/S values of the dyed fabrics, to determine the most appropriate dyeing conditions with mordant.

The extract of Coreopsis tinctoria can dye silk fabrics satisfactorily. Good dyeing depth and fastness can be obtained by using suitable dyeing methods and dyeing conditions, especially when using the natural mordant pomegranate rind and the rare earth mordant neodymium oxide. The silk fabrics dyed with Coreopsis tinctoria have good UV resistance, which allows a desirable finishing effect to be achieved while dyeing, using a safe and environmentally friendly method.

The composition of Coreopsis tinctoria is complex, and the specific composition of colouring the silk fibre has not been determined. There are many factors that affect the dyeing experiment, which have an impact on the experimental results.

The results of this study may help expand the application of Coreopsis tinctoria beyond medicine.

To the best of the authors’ knowledge, this paper is the first report on dyeing silk with the extract of Coreopsis tinctoria achieving good dyeing results. Its depth of staining and staining fastness were satisfactory. Optimum dyeing method and dyeing conditions have been identified. The fabric dyed with Coreopsis tinctoria has good UV protection effect, which is conducive to improving the application value of the dyeing fabric. The findings help offer a new direction for the application of medicinal plants in the eco-friendly dyeing of silk.

This paper aims to analyze a failure case of a P110 tube in a CO₂ flooding well.

The chemical composition, microstructure and mechanical properties of the failed P110 tubing steel were tested, and met the API Spec 5CT standard. The fractures were investigated by scanning electron microscopy and energy dispersive spectroscopy.

Fracture was induced by stress corrosion cracking (SCC) and that the stress concentration caused by the mechanical damage played an important role in the failure. The failure case is a SCC failure affected by mechanical damage and galvanic corrosion.

The effect of the infiltration of groundwater was studied in the failure case. The stress concentration caused by the mechanical damage played an important role in the failure.

The purpose of this paper is to investigate the effects of parameters on the liquid phase migration (LPM) during the freeze-form extrusion fabrication (FEF) process.

To carry out this study, three factors were systematically investigated using orthogonal design of experiments. These three parameters are the extrusion velocity, the extrusion interval time and the extrusion head length. An orthogonal array with nine test units was selected for the experiments. Range analysis and analysis of variance were used to analyze the data obtained by the orthogonal experiments to identify the order of significant factors on LPM.

It was found that the LPM decreased with the increase of extrusion velocity and increased with the lengthening of extrusion interval time and the length of the extrusion nozzle. The order of significant factors for the LPM were found to be extrusion velocity > extrusion nozzle length > extrusion interval time.

Using an orthogonal design of experiments and a statistical analysis method, the liquid content of extrudate can be predicted and appropriate process parameter values can be selected. This leads to the minimization of LPM during the FEF process. Also, this analysis method could be used to study the LPM in other paste extrusion processes.

This paper suggests that the factors have significant impact on LPM during FEF process. The following analysis in this paper is useful for FEF users when prediction of LPM is needed. This methodology could be easily applied to different materials and initial conditions for optimization of other FEF-type processes. The research can also help to get better understanding of LPM during the FEF process.

The purpose of this paper is to study the functionality of additively manufactured (AM) parts, mainly depending on their dimensional accuracy and surface finish. However, the products manufactured using AM usually suffer from defects like roughness or uneven surfaces. This paper discusses the various surface quality improvement techniques, including how to reduce surface defects, surface roughness and dimensional accuracy of AM parts.

There are many different types of popular AM methods. Unfortunately, these AM methods are susceptible to different kinds of surface defects in the product. As a result, pre- and postprocessing efforts and control of various AM process parameters are needed to improve the surface quality and reduce surface roughness.

In this paper, the various surface quality improvement methods are categorized based on the type of materials, working principles of AM and types of finishing processes. They have been divided into chemical, thermal, mechanical and hybrid-based categories.

The review has evaluated the possibility of various surface finishing methods for enhancing the surface quality of AM parts. It has also discussed the research perspective of these methods for surface finishing of AM parts at micro- to nanolevel surface roughness and better dimensional accuracy.

This paper represents a comprehensive review of surface quality improvement methods for both metals and polymer-based AM parts.