Search results

The aggregation of information from a group of decision experts for developing collective opinion is the important question in practice. The purpose of this paper is to provide a group decision-making method via ordered weighted aggregation (OWA) operator and grey incidence analysis.

In this study, OWA operator provides aggregation of attribute values to form an overall decision for each decision expert, and grey incidence model provides aggregation of decision experts’ evaluations to form overall score for each alternative. The example illustrates the procedure and practicability of the proposed model.

A new thought for multiple attribute group decision-making problems is given. The proposed method produces an overall desirability score for each alternative.

This is to obtain a more comprehensive and realistic solution to the given group decision-making problem. The proposed analysis method of group decision-making problems reveals vitality of grey systems theory.

This paper combines OWA operator and grey incidence analysis to obtain a novel and effective method for group decision making. It is suitable for group decision-making problems in which the attribute weights are completely unknown, expert weights are completely unknown.

This study aims to minimize the warpage issue in memory-computing integrated chiplets with 2.5D packaging on a large-scale wafer subjected to multistress through synergistic optimization of key structure paraeters.

In this study, memory-computing integrated chiplet-based 2.5D packaging was designed and the warpage optimization under electro-thermal-vibration coupling was conducted with finite element analysis simulation. Compared studies were also conducted with the imposing condition of single electrical and thermal stress.

The research results indicated that electrical and thermal stress had a significant impact on the packaging warpage while that of vibration on warpage was minimal. For structure parameters, the chiplet thickness had a significant effect on the warpage of the model, while the influence of chiplet size was relatively small. When the chiplet thickness decreased to 100 µm, the warpage was reduced by 8.96%. Under thermal stress loading, the impact of packaging density on the overall warpage of the chiplet-based 2.5D packaging model is relatively small. However, under the loading of electrical stress or electro-thermal coupling, the packaging density has a severe impact on warpage, with the maximum approaching 1.3011 µm for just one chiplet. Compared with vibration alone, electro-thermal-vibration coupling slightly increased the warpage, which is primarily evident in the exacerbation of warpage in stacked chiplets and edge locations. Lower packaging density with 150 µm chiplet thickness contributed to a minimal warpage in stacked chiplet, the edge locations of which exhibited a relatively severe warpage.

The research provides a theoretical basis for warpage optimization of memory-computing integrated chiplets with 2.5D packaging subjected to multistress coupling.

When designing the layout of chips for large-sized high-reliability silicon substrates, this method can be used to control the most severely warped areas at the edges, thus making it easier to optimize warpage.

With the approaching of physical limits dictated by Moore’s Law, advanced packaging technologies, primarily centered around 2.5D packaging and three-dimensional packaging, were attached importance to. It can integrated multifunctional chips, which have become increasingly complex, leading to greater difficulty in design and implementation, and an obvious increase in overall manufacturing costs. In this context, chiplets offered a viable approach for future chip designs. With the high demand for high-performance computing, memory-computing integrated chiplets were designed. The warpage of 2.5D packaging by altering structural parameters under multiphysical field coupling conditions was studied to provide a theoretical basis for warpage optimization of memory-computing integrated chiplets with 2.5D packaging.

With the development of science and technology, more creators are trying to use new crafts to represent the cultural trends of the social media era, which makes cultural heritage innovative and new genres emerge. This compels the academic community to examine craft from a new perspective. It is very helpful to understand the hidden representational structure of craft more deeply and improve the craft innovation system of cultural and creative products that we deconstruct the craft based on Complex Network and discover its intrinsic connections.

The research crawled and cleaned the craft information of the top 20% products on the Forbidden City’s cultural and creative products online and then performed Complex Network modeling, constructed three craft representation networks among function, material and technique, quantified and analyzed the inner connections and network structure of the craft elements, and then analyzed the cultural inheritance and innovation embedded in the craft representation networks.

The three dichotomous craft representation networks constructed by combining function, material and technique: (1) the network density is low and none of them has small-world characteristics, indicating that the innovative heritage of the craft elements in the Forbidden City’s cultural and creative products is at the stage of continuous exploration and development, and multiple coupling innovation is still insufficient; (2) all have scale-free characteristics and there is still a certain degree of community structure within each network, indicating that the coupling innovation of craft elements of the Forbidden City’s cultural and creative products is seriously uneven, with some specific “grammatical combinations” and an Island Effect in the network structure; (3) the craft elements with high network centrality emphasize the characteristics of decorative culture and design for the masses, as well as the pursuit of production efficiency and economic benefits, which represent the aesthetic purport of contemporary Chinese society and the ideological trend of production and life.

The Forbidden City’s cultural and creative products should continue to develop and enrich the multi-coupling innovation of craft elements, clarify and continue their own brand unique craft genes, and make full use of the network important nodes role.

BIM technology has a huge potential for improving the renovation efficiency for as-built buildings. However, due to the absence of raw design drawings and the complex interior environment, it is difficult to implement 3D reconstruction of building interiors in interior renovation projects. Therefore, this study proposes a 3D reconstruction framework of building interiors, with an aim to generate building interiors building information modeling (BIM) models quickly and accurately based on scan-to-BIM and generative design.

The proposed framework begins by reconstructing interior structured elements based on the scan-to-BIM process including collecting accurate information of as-built buildings by laser scanning, obtaining point clouds of structured elements through deep learning and developing an efficient dynamo algorithm workflow for generating structured elements BIM model. For unstructured elements, intelligent layout design and efficient BIM generation are conducted by combining the BIM tools and generative design.

The successful implementation of the proposed framework in a conference room demonstrated the feasibility of the proposed framework. The semantic segmentation scheme based on deep learning also exhibited excellent recognition and high efficiency for interior structured elements. Furthermore, it is proved that the combination of scan-to-BIM and generative design has high application value in the 3D reconstruction of building interiors.

On one hand, a feasible framework is proposed to generate BIM model of building interiors, improve interoperability among different software tools, streamline the complexity of the scan-to-BIM process and meet the reconfiguration requirement of unstructured elements layout scheme in interior renovation projects. On the other hand, the use of BIM and various emerging technologies can drive digital transformation and further advance the industrialization process of interior renovation in as-built buildings.

The purpose of this paper is to explore the potential challenges in developing numerical methodologies for inverse problems and optimizations.

Summarizing previous research results mainly contributed by two research groups of Zhejiang University and Hong Kong Polytechnic University.

Computational intelligence plays an essential role in studying inverse problems and optimizations.

An up-to-date review on the current status of numerical methodologies, especially computational intelligences, for inverse problems and optimizations contributed by Chinese researchers.

Human assembly process recognition in human–robot collaboration (HRC) has been studied recently. However, most research works do not cover high-precision and long-timespan sub-assembly recognition. Hence this paper aims to deal with this problem.

To deal with the above-mentioned problem, the authors propose a 3D long-term recurrent convolutional networks (LRCN) by combining 3D convolutional neural networks (CNN) with long short-term memory (LSTM). 3D CNN behaves well in human action recognition. But when it comes to human sub-assembly recognition, the accuracy of 3D CNN is very low and the number of model parameters is huge, which limits its application in human sub-assembly recognition. Meanwhile, LSTM has the incomparable superiority of long-time memory and time dimensionality compression ability. Hence, by combining 3D CNN with LSTM, the new approach can greatly improve the recognition accuracy and reduce the number of model parameters.

Experiments were performed to validate the proposed method and preferable results have been obtained, where the recognition accuracy increases from 82% to 99%, recall ratio increases from 95% to 100% and the number of model parameters is reduced more than 8 times.

The authors focus on a new problem of high-precision and long-timespan sub-assembly recognition in the area of human assembly process recognition. Then, the 3D LRCN method is a new method with high-precision and long-timespan recognition ability for human sub-assembly recognition compared to 3D CNN method. It is extraordinarily valuable for the robot in HRC. It can help the robot understand what the sub-assembly human cooperator has done in HRC.