Aiming at the problem of the current common Convolutional Neural Network(CNN)in the expression and gender recognition tasks,that is training process is complicated,time-consuming,and poor in real-time performance,a realtime facial expression and gender recognition model based on depthwise separable convolutional neural network was proposed. Firstly,the Multi-Task Convolutional Neural Network(MTCNN)was used to detect faces in different scale input images,and the detected face positions were tracked by Kernelized Correlation Filter(KCF)to increase the detection speed. Then,the bottleneck layers of convolution kernels of different scales were set,the kernel convolution units were formed by the feature fusion method of channel combination,the diversified features were extracted by the depthwise separable convolutional neural network with residual blocks and separable convolution units,and the number of parameters was reduced to lightweight the model structure. Besides,real-time enabled backpropagation visualization was used to reveal the dynamic changes of the weights and characteristics of learning. Finally,the two networks of expression recognition and gender recognition were combined in parallel to realize real-time recognition of expression and gender. Experimental results show that the proposed network model has a recognition rate of 73. 8% on the FER-2013 dataset,a recognition rate of 96% on the CK+ dataset,the accuracy of gender classification on the IMDB dataset reaches 96%;and this model has the overall processing speed reached 70 frames per second,which is improved by 1. 5 times compared with the method of common convolutional neural network combined with support vector machine. Therefore,for datasets with large differences in quantity,resolution and size,the proposed network model has fast detection,short training time,simple feature extraction, and high recognition rate and real-time performance.

Aiming at the problem of the dependence between software system modules, a hybrid defect prediction model based on network representation learning was constructed by analyzing the network structure of software system. Firstly, the software system was converted into a software network on a module-by-module basis. Then, network representation technique was used to perform the unsupervised learning on the system structural feature of each module in software network. Finally, the system structural features and the semantic features learned by the convolutional neural network were combined to construct a hybrid defect prediction model. The experimental results show that the hybrid defect prediction model has better defect prediction effects in three open source softwares, poi, lucene and synapse of Apache, and its F1 index is respectively 3.8%, 1.0%, 4.1% higher than that of the optimal model based on Convolutional Neural Network (CNN). Software network structure feature analysis provides an effective research thought for the construction of defect prediction model.

Since the heterogeneous multi-agent systems composed of first-order and second-order agents have input and velocity saturation characteristics, the systems cannot achieve consensus. A leaderless distributed controller and a leader-following distributed controller of the heterogeneous multi-agent systems were constructed, and the sufficient conditions ensuring the successful consensus were given. The range of communication gain was calculated by using Lyapunov stability theory and Lasalle invariant set principle. The result of numerical simulation of leaderless heterogeneous multi-agent systems with input and velocity saturation characteristics shows that when the communication gain is not in a proper range, consensus of the leaderless heterogeneous multi-agent systems can not be achieved. However, when the proposed gain selection method is used to select the communication gain, the heterogeneous multi-agent systems can overcome input and velocity saturation characteristics, then consensus of the leaderless heterogeneous multi-agent systems can be achieved. The result of numerical simulation of leader following heterogeneous multi-agent systems with input and velocity saturation characteristics proves that the calculation method of the communication gain is also applicable to the semi-global leader-following consensus control of the heterogeneous multi-agent systems.

Concerning the stationary group consensus problem for the heterogeneous multi-Agent systems, which are composed of first-order Agents and second-order Agents, two stationary group consensus protocols were proposed under fixed interconnection topology and switching interconnection topologies respectively. By constructing Lyapunov-Krasovskii functions, the sufficient conditions, which are formulated as linear matrix inequalities, were obtained for the system converging to the group consensus asymptotically under the group consensus algorithm with identical time-varying communication delay. Finally, the simulation results show that the heterogeneous multi-Agent systems with time delay converg to the group consensus asymptotically under certain conditions.

The existing detection approaches for fake plate vehicles have high cost and low efficiency. A new parallel detection approach, called TP-Finder, was proposed based on historical Automatic Number Plate Recognition (ANPR) dataset. To effectively handle the data skew problem emerged in the parallel processing of large-scale dataset, a new data partition strategy based on the idea of integer partition was implemented, which obviously improved the performance of fake plate vehicle detection. Besides, a prototype system for recognizing fake plate vehicles was developed based on the TP-Finder approach, and it could exactly present historical trajectories of all suspicious fake plate vehicles. Finally, the performance of TP-Finder approach was verified on a real ANPR dataset from a city. The experimental results prove that the partition strategy of TP-Finder can achieve a maximum of 20% performance improvement compared with the default MapReduce partition strategy.

There are few works performed on the application and analysis of the Fork/Join framework at present. This paper refactored several benchmarks, including series, crypt, sparsematmult and sor, in the Java Grande Forum (JGF) benchmark suite by using Fork/Join framework. Taking the series benchmark as an example, detailed refactoring process was presented. In the experimentation, the execution time of each benchmark was evaluated with different threshold, and the execution time of Fork/Join framework was compared with that of multi-threaded version. Furthermore, the number of work-stealing operations was presented. The experimental results show that the execution time using Fork/Join framework can reduce 14.2% on average than that using multi-thread. When evaluating the series benchmark with data size sizeC and two threads, the execution time of Fork/Join framework is 40% lower than that with multi-thread. Programs using the Fork/Join framework can get better performance than that using multi-thread.

In order to overcome the demerits of poor deeply searching ability and easily relapsing into local extremum in basic Fruit Fly Optimization Algorithm (FOA), a new algorithm named Shuffled Fruit Fly Optimization Algorithm with Local Deep Search (SFOALDS) was proposed. The local optimal individual in each group was deeply searched circularly by referencing updating strategy of Shuffled Frog Leaping Algorithm (SFLA). SFOALDS not only efficiently avoids relapsing into local extremum, but also improves convergence velocity and convergence precision in the late evolution. The experimental results show that the proposed algorithm has better global searching performance than basic FOA and SFLA, especially on high dimensional functions.

In order to overcome the problems of low convergence precision and easily relapsing into local extremum in basic Fruit Fly Optimization Algorithm(FOA), by introducing the chaos algorithm into the evolutionary process of basic FOA, an improved FOA called Adaptive Chaos FOA (ACFOA)is proposed. In the condition of local convergence, chaos algorithm is applied to search the global optimum in the outside space of convergent area and to jump out of local extremum and continue to optimize. Experimental results show that the new algorithm has the advantages of better global searching ability, speeder convergence and more precise convergence.

In this paper, attraction and exclusion operations of bacterial chemotaxis were introduced into original Fruit Fly Optimization Algorithm (FOA), and FOA based on Bacterial Chemotaxis (BCFOA) was proposed. Exclusion (to escape the worst individual) or attraction (to be attracted by the best individual) was decided to perform by judging the fitness variance is zero or no, so that the problem of premature convergence caused by the loss of population diversity, which resulted from the fact that individuals only were attracted by the best one in FOA, was solved. The experimental results show that the new algorithm has the advantages of better global searching ability, and faster and more precise convergence.

With the rapid urbanization and industrialization, the risk of urban disasters increases and the layout of the city fire stations cannot meet fire safety requirements. In order to effectively overcome the traditional polygon coverage and edge coverage in the location problem of big error and low efficiency, the algorithm expanded the discrete (node-edge) graphic patterns as the continuous network structure to achieve the continuous coverage of road network. In the optimization of the selection process, this paper introduced the simulated annealing and improved the annealing process. A case has verified the selection method of location, and a comparative analysis of the coverage indicators of the location between discrete and continuous optimization was made. The result shows that the continuous optimization has better performance. It is of feasibility, generality and rationality and is a preferable method to resolve the site selection of the urban fire station properly.

Since the current process performance measurement software for recycling enterprise was lack of practical models and appropriate methodology to direct the identification and selection of process performance indicators, this paper firstly brought forward a formal model of recycling enterprise process performance indicators selection based on fuzzy covering set, which highlighted the alignment with enterprise strategy. Then, under the guidance of Balanced Scorecard philosophy, this paper constructed an instantiated model of Recycling Enterprise Process Performance Indicators Selection with Strategic Alignment (SA-REP2IS), which can correlate the strategy of recycling enterprise and the related performance indicators. A corresponding problem-solving strategy for the model was also proposed. Finally, a case study was given to illustrate the application of the above proposed methods.