Design for testability is one of the important research interests in software engineering. It becomes crucial in model-based testing because software behavioral models can be used to construct test sequences to perform conformance testing. However, testability of models has received less attention in the past. A model with high testability is easy to be used to construct effective test sequences, and conformance testing can also be realized easily. To improve testability of software behavioral models, using the formal method, we present five testability metrics: observability, controllability, test constructibility, performability, and error traceability. Then, a case is studied to evaluate the effectiveness of proposed testability metrics. As a result of the case study, models with higher testability can not only be used to generate executable test sequences, but also the size of the constructed test suite is also smaller. Our research will enrich the modeling theory of model-based testing and can improve the application of this test method in industry.

Submitted on October 29, 2017; Revised on November 12, 2017; Accepted on December 3, 2017
References: 34

Software testing is an important means to ensure the quality of software. However, there is a lack of effective modeling and optimization of CPS software testing. In this paper, Petri nets are used to model the underlying devices, components, connectors and test cases of CPS software. Aspect oriented programming extracts the crosscutting concerns of CPS software testing. The behaviors and their relationships are described based on AOP, and a weaving mechanism is used to dynamically integrate these models into the test model of CPS. The correctness of the model is analyzed by using the operation semantics and related theories of Petri nets. Based on the state space of constructed model, a strategy is proposed to dynamically select the test suite. The experiment results show that this method can effectively describe the CPS software testing process, which can improve the quality of software testing.

Submitted on September 29, 2017; Revised on November 12, 2017; Accepted on November 23, 2017
References: 17

Human-Machine Interface (HMI) of Computer Numerical Control (CNC) machine control panel affects the work efficiency and experience of users. For many operators, the CNC machine control panel causes a heavy burden for them because it is not easy to use. The designer creates the shape, color and layout of CNC machine control panel to meet the user’s needs. There is some design schemes produced for managers or designers to make choices. Effective design evaluation can help the designer find usability issues of human-machine interface and iterative optimization schemes. In this study, a practical exploration relating to HMI of CNC machine control panel was carried out using the multidimensional experimental data synchronous testing analysis method. Four HMI design schemes of CNC machine control panel were evaluated. The physiological measurement instruments, eye tracker and behavior analyzer were adopted to obtain the user’s physiological data, psychological data and behavioral data. Scientific statistical results were formed and output by visualization. The designer can choose the best schemes and find the design elements to be improved by this mutually validating method. It’s helpful for the designer to improve the ergonomics of human-machine interface design of CNC machine control panel.

Submitted on September 20, 2017; Revised on November 10, 2017; Accepted on November 30, 2017
References: 35

Cloud testing, with the features of automatic deployment, parallel submission, on-demand distribution and timely response, has been widely favored by many users. Therefore, it is crucial to reduce energy consumption, satisfy user requirement and timely response to user requests for resources, which are guaranteed by a good resource scheduling scheme. The requirements and benefits between user and provider of cloud testing are comprehensively measured in this work. On one hand, in order to meet the expectations of different users for the finish time and cost of their tasks, the definition of user expectation is introduced and then a dynamic pricing model is constructed to achieve the flexible conversion between time and cost. On the other hand, genetic algorithm is employed to implement resource scheduling in cloud testing, which can shorten the running time of all tasks on the cloud testing platform to improve the efficiency and reduce the load as greatly as possible. Finally, comparative experiments show that the scheme proposed in this work is feasible and efficient.

Submitted on October 8, 2017; Revised on November 6, 2017; Accepted on November 23, 2017
References: 29

Safety-Critical Cyber-Physical Systems (SCCPSs) are a special kind of Cyber-Physical Systems (CPSs) which highlight the importance of system correctness and safety. To apply automatic testing or model checking technique in CPSs, a model that fully captures the features is required to serve as input. So, a novel efficient spatio-temporal language and the analysis techniques are demanded to support both temporal and spatial expression and reasoning. In fact, a synchronous language, LUSTRE, is widely used in safety-critical systems development. However, LUSTRE lacks spatial constructors. Thus, it is difficult to express the behaviors related to spatial features in SCCPSs. In this paper, we propose a language named ST-LUSTRE to support the unified modeling of spatial and temporal properties of CPSs. We define the syntax and semantics of ST-LUSTRE. Its semantics is interpreted on the topological space and natural number which is based on time sets. We also specify typical SCCPSs properties in ST-LUSTRE. ST-LUSTRE is successfully applied to a communication based train control system of Shanghai Fuxin Intelligent Transportation Solutions CO.,Ltd. (FITSCO).

Submitted on October 11, 2017; Revised on November 9, 2017; Accepted on November 25, 2017
References: 17

This paper proposes an accurate channel calibration algorithm for ground moving target indication (GMTI) applications with multichannel synthetic aperture radar (SAR) systems. The proposed algorithm is developed in a statistical fashion, and a useful tool called correlation analysis is employed, which measures the coherence between two channels by means of correlation coefficient. Based on correlation analysis theory, an optimum solution to a very general channel calibration problem is obtained, for which only the knowledge of the first two moments of the channel output signals is assumed and no other distributional assumptions are made. The proposed optimum solution represents a generic solution to the general channel calibration problem, which leads to a simple implementation form and needs no iterative operation. The performance test results show that the proposed correlation analysis processing algorithm is an effective way of calibrating the channels.

Submitted on October 29, 2017; Revised on November 12, 2017; Accepted on December 1, 2017
References: 21

Recommender systems play an important role in e-commerce. This paper discusses three classical methods - offline analytics, user study, and online experiment - to evaluate the performance of recommender systems and also analyzes their application scenarios. Some performance evaluation metrics of recommender systems are reviewed and summarized from four perspectives (machine learning, information retrieval, human-computer interaction and software engineering) combined with the above three evaluation methods. These evaluation methods and evaluation metrics summarized in the paper provide the designers with guidance for the comprehensive evaluation and selection of recommended algorithms.

Submitted on October 29, 2017; Revised on November 17, 2017; Accepted on December 1, 2017
References: 32

With the intensive and large-scale development of cloud computing, software testing has become one of the most important problems. How to evaluate and assess the testing process of a cloud computing system is a key to the deployment and use of it. This paper proposes a method to generate and optimize test paths of cloud application based on SDN, which separates the cloud service testing process from the underlying execution logic, thus improving the scalability of the testing process. The finite state machine (FSM) is used to establish the formal description language of the cloud application testing process, and is also used to model the basic elements, such as cloud services, jobs, test cases and cloud applications, and construct a test model for cloud application. The related theory of FSM is used to analyze the effectiveness and correctness of cloud application test model. Based on the actual mapping of the model, the generation method of test path is also given. In addition, we analyze the coverage of test cases and propose the test path optimization methods to improve the test efficiency. The specific examples and simulations show that this method can simplify the design and analysis of the cloud application testing process and effectively improve the generation of test paths.

Submitted on September 30, 2017; Revised on November 1, 2017; Accepted on November 17, 2017
References: 26

With the intelligent development of digital government management services and the advancement of Knowledge Graph study, it is necessary and possible to construct and verify a sound knowledge base of political and economic news to satisfy the users’ requirement of learning the information. Due to the high profession and diversity of political and economic news data, the entity link in the initially constructed knowledge base is lacking completeness. Meanwhile, the high frequency of data update leads to the iterative update of knowledge base. To address the problems, this paper builds a comparatively effective system in which we apply the reasoning results to the construction and iterative update of the political and economic news knowledge base. Then, a syncretic reasoning algorithm based on Subgraph Feature Extraction (SFE) and the factorization of a three-way tensor (RESCAL) is proposed to predict a link and accomplish the reasoning. Using the field data of political and economic news as a case of engineering application, the system we built effectively solves the incompleteness of the entity link in the initial knowledge base and the iterative update problem. The function of knowledge reasoning module and iteration module of knowledge base construction and autonomous updating system are verified by designing and implementing knowledge reasoning, as well as updating knowledge iteration. The experimental results demonstrate the effectiveness and feasibility of the functions of the knowledge base construction and autonomous updating system are verified.

Submitted on October 29, 2017; Revised on November 12, 2017; Accepted on December 3, 2017
References: 30

Genetic Algorithm (GA for short), which simulates the process of natural evolution to search and achieve the optimal solution, is often employed to generate test cases. Therefore, a GA-based test case generation policy, which introduces the concept of node probability coverage as the detection method for nodes in unreachable paths, is proposed in this work. Moreover, in the application under test, complex decisions and nested structures often lead to different execution difficulty of each statement. Therefore, a path coverage method based on contact layer proximity is presented, which quantifies the difficulty difference of different statements using contact vector. Besides, contact layer proximity and node probability coverage are combined to design the fitness function in GA. Then, the experiments about two classical benchmark cases, namely triangle-classifying program and bubble sort program, are conducted. The result is compared and analyzed with a similar method, namely the method of node probability coverage. It is shown that the proposed test case generation method is more efficient. Finally, a plug-in using the proposed test case generation method is developed.

Disambiguating named entities is an important problem in natural language processing, knowledge base, question answering systems. In the paper, we propose a Markov logic network knowledge graph solution for solving entity resolution problem. First, we employ knowledge graph to represent the entity relationship between linked entities in the knowledge base. Then, we utilize MLN to inference the inconsistent relationship in the knowledge graph, and disambiguate the entities in the process of entity disambiguation. As far as we know, inferencing with MLN is a first attempt for entity disambiguation in the knowledge graph. We evaluate the proposed solution with three real world knowledge bases and compare it with four baseline solutions. The experimental results demonstrate that our solution is 7% higher than other baseline methods with F1 measure. We also test our scheme and compare entity resolution systems on four datasets with three knowledge base corpora. Extensive experiments show that our solution achieves higher precision and recall than baseline solutions.

Submitted on July 25, 2017; Revised on August 30, 2017; Accepted on September 15, 2017(This paper was presented at the Third International Symposium on System and Software Reliability.
References: 30

In order to keep the motor volume and slot number unchanged as well as reduce the motor speed, a new type of linear permanent magnet vernier motor (LPMVM) is discussed. The characteristics of LPMVM are simple structure, low speed, and reduced thrust ripple. According to the air-gap permeance about slots only on the stator and the formulas of flux density with no-load EMF, the thrust are analyzed, and the motor structure is given. Its steady performances, compared with the traditional permanent magnet linear synchronous motor (PMLSM), are simulated by using the finite element method (FEM) as compared with the traditional permanent magnet linear synchronous motor (PMLSM). The results show that it has a good performance in a low speed, especially since the detent force is very small, so it is suitable for the low-speed direct-drive system.

Submitted on July 25, 2017; Revised on August 30, 2017; Accepted on September 15, 2017(This paper was presented at the Third International Symposium on System and Software Reliability.
References: 12

PADS (Process Algebra for Demand and Supply) is a formal framework to analyze hierarchical scheduling in real-time embedded systems. Inspired by the supply simulation relation in PADS, we introduce a partial supply simulation relation in order to describe the fact that an unschedulable task may finish on time. It is more general than the supply simulation relation. Then, we explore some properties of partial supply simulation relation. Furthermore, we establish a proof system for the partial supply simulation relation in a decomposing-composing way, which helps to infer tasks' partial schedulabilities. Finally, it is proved that the proof system is sound and complete with respect to the semantic definition of partial supply simulation relation.

Submitted on July 25, 2017; Revised on August 30, 2017; Accepted on September 15, 2017(This paper was presented at the Third International Symposium on System and Software Reliability.
References: 18

High Efficiency Video Coding (HEVC) possesses a special intact coding flow frame that involves four basic procedures: prediction, transformation, quantization and entropy coding. The whole process has high computational complexity, especially the transformation and quantization, which obstruct HEVC for real-time application. To reduce the computational complexity, computing time, or resources, a fast rate-distortion optimized algorithm for HEVC is presented. The rate distortion optimized transformation (RDOT) with reclassification and the rate-distortion optimized quantization (RDOQ) extensional bypass decision method are employed to save the computational complexity. The RDOT performs initial classification and reclassification for each class to simplify the decorrelation of residual data transformation. Moreover, the RDOQ bypass could reduce the computational complexity by skipping the unnecessary candidates of transformation blocks with high computation. Experimental results reveal that the proposed algorithm implemented reduces HEVC coding complexity and achieves almost the same coding performance.
(This paper was presented at the Third International Symposium on System|RevSoftware Reliability.
Submitted on July 25, 2017; Revised on August 30, 2017; Accepted on September 15, 2017(This paper was presented at the Third International Symposium on System and Software Reliability.
References: 17

With the deepening of the integration of information technology and industrialization, industrial control systems of computerized numerical control (CNC) machines is gradually changed from the original isolated closed mode into the Internet model. It is not only facing the internal threat, but also facing the threat from the Internet. The existing industrial control system of CNC machine is due to long-term in a closed environment. The system cannot be updated in time; it is difficult to defend against the threat of industrial networks from the Internet. In view of the above problems, this paper firstly puts forward the security and trusted framework for CNC machine control system based on trusted computing technology, and elaborates the frame composition and function principle in detail. Then, the module of the trusted communication monitoring and control for the control system of CNC machine is presented; it can realize the scalability of the CNC machine system and the correlation of the equipment, while satisfying the trusted measurement function. Finally, this paper analyzes the reliability of the traditional CNC machines by means of experimental simulation. The performance shows that in the controllable range, the proposed framework can effectively enhance the CNC machine computing environment security.

Submitted on September 29, 2017; Revised on November 12, 2017; Accepted on November 23, 2017
References: 16

This paper proposes an optimal path planning method for the multiple automated guided vehicle (AGV) system based on two-staged scheduling; at the offline scheduling stage, high degree of genetic algorithm is used for the optimal obstacle avoidance path planning of AGV under the static environment, which cannot only solve the premature convergence of genetic algorithm, but also the obstacle avoidance of AGV path planning. Online scheduling stage mainly refers to test the node conflict, opposite conflict and pursuit conflict between AGV and these conflicts are solved to achieve online collision avoidance scheduling for AGV. Finally, the paper uses the secondary developed openTCS for algorithm simulation. The processing methods when all kinds of conflicts occur are simulated in the multi-AGV systems, and the results show that the method is effective and reliable for the path planning of multi-AGV systems.

Submitted on October 11, 2017; Revised on November 12, 2017; Accepted on November 23, 2017
References: 15

For traditional computer-aided diagnosis, the feature extraction of lung nodules only relies on artificial design, and the use of morphological features may lose nodular information, causing differences in classification results. This paper proposes a method of lung nodule feature extraction and classification as benign or malignant. First, the region of interest (ROI) of lung nodules was obtained from the original CT image using a threshold probability map. Next, the deep features of the lung nodules were extracted using the deep belief network (DBN). Finally, an extreme learning machine (ELM) was used as the classifier for benign and malignant classification. On the publicly available LIDC database, our method reaches a high accuracy of 95±0.3% in the diagnosis of lung nodules, and the area under the ROC curve is 0.932, which is superior to other feature extraction methods. Our method also avoids the complexity of artificial extraction and differences in feature selection, and it can provide a reference for clinical diagnosis.

Submitted on October 20, 2017; Revised on November 18, 2017; Accepted on November 29, 2017
References: 24

To address the deficiencies of cryptography education, we design and implement a virtual experiment system named VESC (Virtual Experiment System for Cryptography), which is composed of a front end developed with HTML5 and a back end developed with Node.js and Docker. Based on the parameters input by the users, VESC uses the cross-browser vector graphics library Raphael to dynamically demonstrate the workflow of various cryptography algorithms to users in a step-by-step manner. VESC is easy to operate while having a friendly interface. After users submit their codes, VESC compiles the codes, executes them, and sends the results back to the users. Our experimental results show that VESC achieves high performance in a highly concurrent accessing environment. VESC is not only a virtual experimental platform for students, but also offers an assistant system for teachers to help students better understand the principles of complicated cryptographic algorithms and protocols as well as their applications.

Submitted on October 1, 2017; Revised on November 2, 2017; Accepted on November 23, 2017
References: 21

In China, new energy is developing rapidly. In recent years, new energy power generation has been installed with explosive growth. However, the coordination problem between new energy penetration capability and the operation mode of the system has not been solved. Especially in the ‘Three North’ areas, new energy is severely limited. As a result, the large-scale new energy medium-term operation optimization algorithm and its parallelization are very urgent. This paper established a mixed integer model for the large-scale new energy medium-term operation problem, and proposed a new method to simplify the 0-1 constraints. Since the most commonly used software has some limitations on solving our mixed integer programming (MIP) problem, we developed a parallel algorithm library (CMIP) V2.0 of our own intellectual-property rights and exploited the parallelism of the algorithm for better performance. Preliminary numerical experiments show that CMIP V2.0 can solve the new energy medium-term operation optimization problem, at least as well as the commercial software CPLEX and the open source software SCIP.

Submitted on October 20, 2017; Revised on November 22, 2017; Accepted on November 30, 2017
References: 23