Advanced Materials Research Vols. 139-141

Abstract: This paper aims to propose a novel three-fold approach to solve dynamic job-shop scheduling problems by artificial immune algorithm. The proposed approach works in three phases. Firstly, priority rules are deployed to decrease problem scale instead of using scheduling algorithms directly. Secondly, immune algorithm is applied to optimize the individual scheduling modules. Finally, integration schema is employed to reschedule operations and minimize makespan of gross schedule. The integration schema is carried out in a dynamic manner that the previous modules’ machine idle time is searched continuously. In this way, the machine utilization is increased while the objective of makespan minimization is maintained. Efficacy of the proposed approach has been tested with test instances of job-shop scheduling problems. The experimentation results clearly show effectiveness of the proposed approach.

Abstract: The ASP pattern's core is to realize on-line rental application software to achieve the win-win of small and medium enterprises (SMEs) and software developer. But the commercial recommendation service is the processing, which treats the massive information, filtrates them to provide users with valuable opportunities, and an improvement on the search strategy. Through the definition model and the solution strategy of the commercial recommendation system, the results are obtained by the collaborative filtering recommendation algorithm. Commercial recommendation service and ASP pattern union, display recommendation system's superiority, thus has a bigger profit.

Abstract: Data is the main subject of enterprise resource planning. It is an important way to enhance economic efficiency and competitiveness, by taking advantage of information technology to implement the data/information management and ensure their objectivity, accuracy and efficiency. In product design and manufacture process, the life time of data resources are divided into production, collection, arrangement, inputting, transmission, using, etc. If data resources information management system and management platform is established, it can record, share and manage data timely and accurately. At the beginning of data management, the clear and reasonable planning is an important foundation to achieve data information management. So, the concrete problems of planning were discussed about management objective, content and mode. In the course of design and manufacture, the data types, data management user and management basic conditions were summarized, and then it was discussed about framework and process of data resource management. Finally, the concrete operation and management measures were discussed on each link of data life cycle. The paper’s idea has a great impact for reference to integration and compromise of product’s data resources, establishment data resources management platform and taking management measures.

Abstract: The real production scheduling problem between steel-making and continuous-casting can be modeled as JSSP with fuzzy processing and delivery time. An improved genetic algorithm is proposed for solving this problem and the improved aspects include the mechanism for preventing early-maturing and the job filter order-based crossover operator. The test results show that the improved genetic algorithm can find better solutions than other three algorithms. A real production scheduling problem of steel-making and continuous-casting is computed using the improved genetic algorithm and it shows the algorithm is effective.

Abstract: The real production scheduling problem between steel-making and continuous-casting can be modeled as Flow Shop Scheduling Problem. Waiting time must be considered for the reason of the temperature decrease of molten steel. The genetic algorithm with an initial population generation mechanism and a piecewise fitness function is proposed for solving the problem. The test results show that the proposed algorithm can find better solutions. A real case of steel-making and continuous-casting scheduling problem is computed and it shows the algorithm is effective.

Abstract: Because of the continuity and discreteness during production, Irradiation industry becomes one of an especial industry. Various product type, uncertain lots and complex processes cycle increase the difficulty of production planning and scheduling. The implementation of CIMS/ERP systems during product processes in the enterprise can not fulfill the special requirements especially in product automation monitoring and quality information monitoring. Research and application of scheduling method and optimization technology effect crucially for enterprise to improve its production efficiency and reduce its production cost. So, more and more scholars pay their attention to this research field. For this reason, the paper presents a mathematical production scheduling model in irradiation workshop based on the research of irradiation industry development and shop scheduling in China and abroad. The author also design an application of the irradiation shop scheduling Optimization with this model in detailed, based on described the concept, principle of genetic algorithm and its method.

Abstract: The training of Support Vector Machine (SVM) is an optimization problem of quadratic programming which can not be applied to the online training in real time applications or time-variant data source. The online algorithms proposed by other researchers have high computational complexity and slow training speed, which can not be well applied to the time-variant problems as well. In this paper the projection gradient and adaptive natural gradient is combined. The constraint projection adaptive natural gradient online algorithm for SVM is proposed. The computation complexity of the constraint projection adaptive natural gradient algorithm is . The learning performance is compared via prediction of the concentration of component A of Continuous Stirred Tank Reactor. The results of simulation demonstrate that the time taken by the constraint projection adaptive natural gradient online algorithm for SVM is far less than that of incremental algorithm, while keep higher prediction precision.

Abstract: This paper uses the BP neural network algorithm to predict the performance of heat exchangers, sets up a applied structure of the BP neural network and expounds the realization of predicted algorithm, including the determination of network structure, the learning rate, the network performance evaluation, the training and test aggregate, the network target errors and the network training times and so on, which is the simulation of predicting the performance of a heat exchanger with pipes buried underground in a ground source heat pump system. The results of prediction show that the relative errors of the heat exchanger performance prediction are mostly within 5.4%, and the neural network prediction results agree well with the experimental results, which have better generalization ability. This research method for underground heat exchanger can provide basis for optimizing the parameters, so it has certain practical significance and social value.

Abstract: The flight simulator is one kind of servo system with uncertainties and disturbances. To obtain high low-velocity performance and good robustness for the flight simulator, we present a novel robust controller based on the acceleration feedback and Disturbance Observer. Firstly, the plant model and the framework of the novel controller are described. Secondly, the principle and the design process of the acceleration feedback controller are analyzed and expatiated respectively. Finally, simulation results on the flight simulator show that the acceleration feedback controller can compensate nonlinear friction problems and the system performance can be improved. Therefore both robustness and high performance of the flight simulator are achieved. It is an applied technology for the control of servo system, such as the flight simulator.

Abstract: Flight motion simulator is one kind of servo system with uncertainties and nonlinearities. To acquire higher frequency response and good robustness for the flight simulator, we present a Backstepping controller based on a Diagonal Recurrent Neural Network (DRNN) to work out this problem. For one thing, the design procedure of the robust Backstepping controller is described. Subsequently, the principle and the design steps of DRNN are analyzed and expatiated respectively. In the end, simulation results on the flight motion simulator show that robust backstepping control based on DRNN can compensate for external disturbances and enhance robustness of the system control performance. Therefore both robustness and high performance of the flight motion simulator are achieved.