This paper presents a joint optimization approach of reliability design (RD) and level of repair analysis (LORA) for fleet systems. A fleet is a multi-machine, multi-indenture system. The present paper investigates the interrelated effect of product design in terms of modularity and inherent reliability with maintenance repair strategy. It proposes a joint approach for the configured fleet system design of reliability and repair decisions at the initial design phase considering the time dependent failure rate of components. The consequences of each design options and level of repair decisions are evaluated based on life cycle cost performance. Additionally, the failure of the machine is modeled using time dependent failure rate models at the part level of the indenture. The methodology integrates many interdependent maintenance decisions such as location of maintenance, type of maintenance (repair/replacement/discard), and indenture level at which maintenance should be performed. The time dependent failure rate of components provides flexibility to consider the effect of practical behavior on the overall fleet level maintenance methodology. This makes the fleet methodology more realistic in terms of optimizing the reliability design and maintenance repair decisions (LORA). This joint problem is the complex combinatorial type problem. To obtain appropriate integrated and disintegrated results for fleet system design and level of repair decisions, genetic algorithm (GA)-based Monte Carlo simulation is used.

In light of trends towards the dual focus on factory design including production line layout and process design and implementation, manufacturing firms are eagerly looking for an integration solution to seamlessly utilize run time performance indices (e.g. throughput) to evaluate the production performance through the design time model (i.e. digital model of system layout). In order to achieve the above goal, this paper introduces a new performance analysis platform (PAP) to close the gap between run time performance measurements and design time model for the low performance (also called performance anti -patterns) detection and refactoring of Smart Production Line. This work expands the above idea in three directions. Firstly, we introduce the design principle of PAP. Secondly, two key models, system layout model specified by UML and system performance model described through Layered Queueing Networks, of the PAP are introduced. A Model-to-Model transformation is presented to transform the design time system model into a performance model for the following performance anti-patterns detection and production line refactoring. A case study shows the early engagement prevents a manufacturer's production system development team from making costly design mistakes while improving the system performance.

In order to fully extract the information that affects the user's arrears and reduce the dimension of features, a hybrid feature selection algorithm based on the particle swarm optimization algorithm with contraction factor (PSOCF) and whale optimization algorithm (WOA), namely, PSOCFWOA is proposed. The PSOCFWOA algorithm combines the advantages of the two algorithms that PSOCF and WOA. The experimental results show that the proposed PSOCFWOA can effectively reduce a large number of redundant or irrelevant features and stably improve the prediction results in the case of low execution time, compared with two state-of-the-art optimization algorithm, and six well-known feature selection approaches to the risk prediction of electricity tariff recovery for power customers.

As for special structure rolling bearing which is not easy to implement a large number of fatigue test, an accurate and efficient reliability calculation method that combines response surface method (RSM) with advanced first order second moment method (AFOSM) was developed under elastohydrodynamic lubrication (EHL). Mechanical model of contact stress analysis was established based on EHL and finite element method (FEM), in which the oil pressure was mapped into hertz contact zone. Considering randomness of correlation factors, limit state function applied to contact fatigue reliability analysis of rolling bearing under EHL was established based on the proposed method and tested by F-test. Compared with traditional Monte Carlo method (MCM), time-consuming of the proposed method only accounts for 0.12％ of MCM, both absolute error of reliability is less than 0.02 and relative error less than 23%. Result shows that the proposed method is accurate and efficient, and could correctly reflect the effect of elliptical contact EHL character on contact fatigue reliability and its sensitivity of rolling bearing.

To reduce the casualty rate in factories, grey system theory (GST) is used to predict safety targets. First, the reliability-based dynamic model and differential equation of grey model GM(1, 1) are established on the basis of GST. Then, the sequence results are accumulated and inverse accumulated further, thus further revealing the change law. According to the analysis results of relative error q, variance ratio C, and small error probability P, the accuracy test for the prediction model indicates that the GM(1, 1) has high prediction reliability and that the prediction precision is a good rank (P > 0.95, C < 0.35). Finally, the prediction results show that the predicted casualty number was 18.2-48.3 in 2017, 35.1-72 in 2018, and 67.9-107.5 in 2019, and both the casualty rate and casualty number will increase over the next three years. The findings in this paper will help guide the establishment of factory safety targets.

Nowadays, there are usually large network groups. If the traditional group key negotiation protocol is adopted, the time cost consumed is very high and unrealistic when the session key calculation requires the participation of members in each group. To solve this problem, we propose a BD group key negotiation protocol based on clustering technology, which solves the defects of parallel key negotiation by clustering. We also conduct a security analysis under the assumption of discrete logarithm problem on elliptic curve and the Diffie-Hellman problem, which theoretically proves its security. Finally, the theoretical analysis and experimental results show that compared with existing key negotiation schemes, this protocol has good robustness, scalability, and lower communication and computing overhead, which has certain guiding significance for the research on user privacy protection in mobile cloud computing networks.

Rotary kilns are widely used in various industrial applications. Most rotary kiln cylinders are supported by three or more stalls. The force state of these rotary kiln cylinders is statically indeterminate, and the calculation process is correspondingly complicated. Aiming at simplifying the calculation model, the structure of the two-stall short kiln based on the design principle of "lateral rigidity and longitudinal flexibility" was designed. The finite element simulation model of the rotary kiln was established using SolidWorks software, and the mechanical behavior analysis of the cylinder was carried out by ANSYS software. The simulation results showed that the deflection value of the cylinder is larger in the vicinity of the kiln and the span, and the load at the bearing is more concentrated. The stress alternates in the rotary kiln body. The stress and strain of the middle area of the rotary kiln cylinder span are very large and close to the allowable stress. The simulation results provide a theoretical reference for the safety assessment and reinforcement design of the cement rotary kiln cylinder.

The performances of current process recommendation need to be improved since they have different degrees of defects. To address this issue, based on metadata, this paper presents a multi-attribute utility group recommendation, which is expected to effectively improve the accuracy of process recommendation. First, a business process description framework (BPDF) is proposed. Then, the similarity between two processes is obtained by calculating the similarity of metadata features. Furthermore, the scenario-oriented group recommendation strategy is developed based on BPDF. The features used in this research are consistent with the key features of the processes in the application. Experimental results show that our approach can improve the effectiveness of business process recommendation.

This paper concerns the relative integrated navigation (RIN) problem for the ship navigation systems with finite horizon energy-limited noises. In the ship RIN systems, two kinds of navigation observations can be utilized to obtain more precise navigation information: the ones sampled by inherent navigation devices of the target ship, and the ones broadcasted by the adjacent ships in the near sea area of the target ship. The latter ones are so called relative navigation observations. In this paper, two kinds of novel sequential finite horizon H∞ fusion filtering algorithms are proposed to deal with these observations. Firstly, a centralized fusion performance index is defined and an augmented centralized H∞ fusion filtering algorithm is given to ensure the performance index. Further, this method is extended as a sequential centralized finite horizon H∞ fusion filtering algorithm to sequentially deal with the relative navigation observations in real time. Then, in the distributed fusion framework, a sequential distributed finite horizon H∞ fusion filtering algorithm is also proposed to fuse the local (relative) state estimates of the target ship. Finally, a simulation is employed to illustrate the validity and feasibility of the proposed sequential finite horizon H∞ fusion filter based relative integrated navigation methods.

Production scheduling is an important decision-making process for manufacturing enterprises. The flexible job shop scheduling problem (FJSP) with worker flexibility is always taken as the dual-resource constrained FJSP (DRCFJSP). However, previous work mainly contributes to the time-related metrics, while ignoring the impacts on the environment. In this paper, an energy-conscious DRCFJSP (ECDRCFJSP) is investigated to optimize the total energy consumption. A decision-making approach is presented based on a modified water wave optimization (MWWO). In the MWWO, three search operators (propagation, refraction, and breaking) are designed following the characteristics of the problem. Extensive simulations are carried out to evaluate the performance of the MWWO algorithm. The computational data demonstrate that our MWWO is effective at solving the considered problem.

This paper investigates the dynamic behavior of half vehicles with a passive suspension system. The study uses an improved pseudo-random road excitation method that has been proposed as a type of road input excitation for vehicles in shock mitigation. For this, five motion equations of the half vehicle model are established, which are derived in terms of four displacements and a pitch angle. Based on the principle of analog equation, the new direct time integration method is used for numerical integration. Both the power spectrum density (PSD) in frequency-domain and the vehicle dynamic response in time-domain against two types of road inputs are investigated and simulated. This includes the PSD of driver vertical acceleration, chassis (vehicle body) vertical acceleration, front (rear) suspension dynamic deflection, front (rear) dynamic tire load, and response of displacement and velocity. The simulation results are achieved for different type of initial inputs. Finally, the value of the specific frequency point of power spectrum density and the time-domain dynamic response of the model are obtained, and the reasons are analyzed.

Signal modulation recognition is the core of cognitive radio and spectrum sensing. With the rapid development and application of deep learning technology in recent years, multi-modal deep learning has become the mainstream of multi-modal machine learning. However, its usage in communication systems has not been well explored. This paper proposes a signal contour stellar images domain recognition method based on deep learning (DL) to achieve the problem of low recognition accuracy under low signal-to-noise ratio. A signal I/Q waveform domain recognition method based on deep complex-valued neural network is proposed to extract the amplitude and phase features of signals to achieve high-precision and high-robustness recognition of multiple signals. A multi-modal deep learning method is proposed to fuse image features, amplitudes, and phase features extracted by complex-valued neural networks to further improve the recognition accuracy and robustness of signals. Finally, the simulation results show the superiority of the scheme and prove that the scheme utilizes the complementarity between signal multi-modalities, removes the redundancy between the modes, and realizes the deep intelligent extraction of signal features, which can lead to a better signal recognition effect.

The Automatic Dependent Surveillance-Broadcast (ADS-B) is a proven technique for aviation control. ADS-B systems periodically transmit the aircraft position, aircraft ID, aircraft velocity and other information on the 1090MHz. Space based ADS-B uses low earth orbit satellites that can achieve global surveillance. A large number of aircrafts will make ADS-B signals overlap more frequently at the receiver, which seriously reduces the decoding efficiency of the receiver. In this paper, we first analyze the collision model, and point out that the one order overlap situation is the most serious in all cases of signal overlap. In this paper, we use the JADE algorithm to separate the one order overlap signals successfully. The experiments show that compared with the traditional decoding algorithm, the decoding accuracy is significantly improved and the decoding efficiency of the receiver is obviously improved after separating the mixed signal.

With the aid of artificial intelligence, this paper builds a machine learning model with the KNN algorithm to optimize the traditional method of urban planning and management (UPM). The optimization is realized through two steps. First, the relevant theories of the updated UPM of livelihood oriented UPM (LOUPM) are explored for the later machine learning architecture design. People's livelihood has a great influence on UPM. Livelihood orientation makes the complex UPM even more complicated due to diverse living needs of citizens. This paper analyzes the deeper relationship between the people's livelihood and UPM, systematically studies the function and connotation of the people's livelihood behavior, and profoundly discusses current contradictions and restraining factors. Second, based on a better understanding of LOUPM, this paper further proposes an artificial intelligence approach to select most related factors to optimize UPM from databases. In the first paper of this series, in the analysis of the relevant UPM theories, three scopes of LOUPM are concluded to be the evaluation of data sets: authority, time, and space. Then, this paper continues to design a software model with the KNN algorithm to evaluate the urban plans and generates optimization advice for the user. Based on this research, it is possible to explore LOUPM. In order to make every effort to seek ways and methods to meet the challenge of the LOUPM in the new period, and by introducing a creative computing approach of government management and social governance, this article can contribute to the growing demand of China's urbanization process.

Obtaining useful information from mass data on the Internet has been a hot topic in information process research in recent years. For un-structural data like online reviews based on natural languages, it becomes more challenging. Online consumer reviews reflect customers' real experience and opinions on products or services. However, there are short of methods or tools to help potential customers find high-quality and helpful reviews from a large number of reviews. This paper applied the concept and idea of creative computing to solve this problem. Tf-idf, as a traditional method to extract text features, measures the importance of words through word frequency and ignores the semantic information in the text data, while the topic model makes up for this deficiency. This paper proposed to use the vector of reviews allocated by LDA topic model to represent text semantic features. Basing on semantic features of reviews, it calculated cosine similarity between the thumb up reviews and other reviews and thus obtain the simulated helpfulness scores of all reviews. Then, a linear regression was designed to obtain two features, i.e., the syntax and semantic features, and determine the simulated helpfulness scores. The proposed method was validated by collected online tourism reviews of Forbidden City and Mount Huang on three Chinese representative online tourism platforms. The results showed that the proposed method can effectively obtain and thus compare the helpfulness of online reviews in a creative way.

Cross-version Software Defect Prediction (CSDP) can be used to predict defect-prone modules or the number of defects in the latest version by using defect rules learned from historical versions. This technique can not only reduce the cost of iterative development of software and improve the reliability of new version, but also be used to understand the causes of defects and improve the software development process. Recently, much effort has been paid to build accurate cross-version defect prediction models, including quality defect predictors, and develop modeling techniques. Liu et al. proposed a new way to construct the predictor based on software code metrics and software process metrics. This new predictor is called Historical Version Sequence of Metrics (HVSM), which is processed with RNN. However, their model has some shortcomings, such as the missing of evolution information in HVSM, difficulty in training RNN, and lack of the ability to predict the number of defects, etc. To solve these problems, we add software evolution metrics to HVSM and bring in a new deep learning technique, Gate Recurrent Unit (GRU), to enhance the HVSM. Our HVSM-based GRU model can predict both defect-prone modules and the number of defects. The experimental results show that the HVSM with incremental evolution metrics provide better performance and in most cases the HVSM-based GRU model outperforms the commonly used baseline models in CSDP.