Applied Mechanics and Materials Vols. 631-632

Abstract: Variable precision rough set (VPRS) based on dominance relation is an extension of traditional rough set by which can handle preference-ordered information flexibly. This paper focuses on the maintenance of approximations in dominance based VPRS when the objects in an information system vary over time. The incremental updating principles are given as inserting or deleting an object, and some experimental evaluations validates the effectiveness of the proposed method.

Abstract: Dominance-based rough sets approach (DRSA) is an effective tool to deal with information with preference-ordered attribute domain. In practice, many information systems may evolve when attribute values are changed. Updating set approximations for these dynamic information systems is a necessary step for further knowledge reduction and decision making in DRSA. The purpose of this paper is to present an incremental approach when the information system alters dynamically with the change of condition attribute values. The updating rules are given with proofs, and the experimental evaluations on UCI data show that the incremental approach outperforms the original non-incremental one.

Abstract: In this paper, a hybrid methodology that incorporates a simulated annealing (SA) approach into the framework of variable neighborhood search (VNS) is proposed to solve the blocking flow shop scheduling problem with the total flow time minimization. The proposed hybrid algorithm adopts SA as the local search method in the third stage of VNS, and uses a perturbation mechanism consisting of three neighborhood operators in VNS to diversify the search. To enhance the intensification search, best-insert operator is adopted to generate the neighbors in SA. To evaluate the performance of the proposed hybrid algorithm, computational experiments and comparisons were conducted on the well-known Taillard’s benchmark problems. The computational results and comparisons validate the effectiveness of the proposed algorithm.

Abstract: Financial markets has witnessed an explosion of algorithmic trading strategy which can help traders especially involved in high-frequency trading efficiently reduce invisible transaction cost. The VWAP strategy usually used by traders can only decrease the cost of price impact by breaking block order into small pieces. However, the behavior of such order splitting may result in inevitable opportunity cost as well as price appreciation. This paper establishes a new algorithmic trading strategy to minimize total transaction costs including price impact, opportunity cost and price appreciation. The results show that the total transaction cost of this optimal trading strategy is lower than VWAP strategy.

Abstract: A permutation flowshop scheduling problem with maximum waiting time constraints to minimize makespan is studied, and a partheno-genetic algorithm (PGA) is presented. In PGA, the fitness function is defined as a decreasing function of makespan to enhance the selected opportunity of good individuals; the roulette algorithm for chromosome selection is improved to keep population diversity with high quality by three strategies: optimal maintenance, fitness adjustment and roulette reconstruction; single-point gene exchange operators are applied to generate offspring. Numerical results demonstrated the feasibility and effectiveness of the algorithm.

Abstract: After years of stable operation, a group enterprise equity management information system has accumulated a large amount of historical operating data. At present, the enterprise operating situation analysis is primarily based on statistical inquiry, so we cannot find out the operational decisions information which includes future trends among the data, and that information are often deeply hidden. In this paper, the decision tree applying to the situation classification prediction is presented, including the construction of a professional field data warehouse, data preprocessing and discretization, decision tree algorithm and tree pruning optimization algorithm using weights. Through the experiment, the prediction accuracy may reach 95.8%. By use of enterprise operating situation analysis based on data mining technology, we can find out useful information for the enterprise management decision, strengthen the correct judgment in marketing operation and other aspects, thereby enhancing the market competitiveness.

Abstract: Digital resource integration is the inevitable trend of library development at present. The author introduces international and domestic representative digital resources integration systems and proposes the goal of digital resource integration. This paper describes the construction process of digital resource integration system based on Metadata storage, including the construction of DC metadata storage, united research system and resource scheduling system.

Abstract: Aiming at the limitations of a single neural network for effective gas load forecasting, a combinational model based on wavelet BP neural network optimized by genetic algorithm is proposed. The problems that traditional BP algorithm converges slowly and easily falls into local minimum are overcame. The wavelet neural network strengthens the function approximation capacity of the network by combining the well time-frequency local feature of wavelet transform with the self-learning ability of neural network. And optimized by the real coded genetic algorithm, the network converges more quick than non-optimized one. This proposed model is applied to daily gas load forecasting for Shanghai and the simulation results indicate that this algorithm has excellent prediction effect.

Abstract: With the development of biology, medical statistics and economy, the study of recurrent event data has made great progress. Many important statistical models of failed time and recurrence process are established. In this paper, we study a joint model based on multiplicable hazard function and proportional hazard function. Considering that the covariate is time-independent, and censoring is dependent on recurrent event processes and end times, we prove the consistency and asymptotic properties of the estimation under certain regularity conditions for this model.

Abstract: Recurrent events data refers to the observation of individuals, which contains the recurrent event time of interest. In this paper, we focus on the statistical analysis of recurrent event, and present a composite model based on the proportional intensity function and additive hazard function, and then prove the consistency and asymptotic properties of the estimation under certain regularity conditions.