Linqing Wang

Genetic algorithm for regionalization problem with adaptive equity constraint

The regionalization problem involves aggregating several spatially contiguous basic geographical units into regions while optimizing a defined objective. Equity is one of its most important constraints with the aim to ensure the value of one or several interesting spatial attributes at a certain level and to provide solutions to specific requirements in applications. This paper tackles a new regionalization scenario in which the threshold used in equity constraint is a regional mean measurement and depends on an unknown amount of regions. A novel nonlinear mixed integer programming model is developed. For the solution of the model a parallel genetic algorithm capable of solving large-scale, real-world instances, is designed. Our efforts in designing a genetic algorithm that integrates an upper bound heuristic are reported. A group of 30 synthetic lattice data is generated according to the spatial auto regressive function to evaluate the proposed genetic algorithm. Equity attributes with both Uniform distribution and Poisson distribution are considered to make a comparison. This work makes an original contribution in the solution of the regionalization problem with adaptive equity constraint.

Subset fusion based T-S fuzzy modeling for blast furnace gas system in steel industry

Blast furnace gas (BFG) is regarded as a very important secondary energy in steel industry, and an effective model to describe the status of BFG system is fairly significant to maintain the system balance and stability. However, the high level noises in industrial data and the disturbances in training samples could lead to the overfitting phenomenon. A fuzzy subset fusion combined with a rule reduction method is proposed in this study to simplify the structure of the rule base and enhance the generalization ability of the fuzzy model. In the proposed method, the parameters of membership functions (MFs) are clustered by using a fuzzy c-means (FCM) method for forming the new representative MFs, and the rules reduction and the consequent parameters update are carried out based on the weights of each rule. The experimental analysis by using a number of real industrial data demonstrates that the proposed method can effectively deal with the fuzzy subset overlapping problem and redundant rules so as to improve the generalization ability of the T-S fuzzy model.

Evolutionary adaptive dynamic programming algorithm for converter gas scheduling of steel industry

It is significant to perform an effective scheduling of byproduct gas system in steel industry for reducing cost and protecting environment. The existing studies largely focused on extracting specific knowledge from human experience or directly optimizing the scheduling performance, which failed to provide a dynamic optimization process for making the scheduling scheme updated online. In this study, an action-dependent heuristic dynamic programming (ADHDP) framework is proposed for the Linz-Donawitz converter gas (LDG) scheduling, in which the scheduling amount is calculated based on the gas system states by utilizing a Tagaki-Sugeno-Kang (TSK) fuzzy model, while a utility function is introduced in the critic network considering the time delay of the gas system to evaluate the scheduling performance over time. For achieving online learning process, the concept of a modified evolutionary algorithm is combined with the ADHDP to obtain the near-optimal scheduling policy at each time instance. To demonstrate the performance of the proposed method, the practical data coming from the energy center of a steel plant are employed. The results show that the proposed method can supply the human operators with effective solution for secure and economically justified optimization of the LDG system.

A dynamic causal diagram and constraint-based method for scheduling in blast furnace gas system of the steel industry

Blast furnace gas (BFG) is one of the important secondary energy in the iron and steel enterprises. The reasonable use of BFG can raise economic profit and alleviate the environment pollution. The existing scheduling mode may cause not only a waste of gas, but also additional operations for compensating the inadequate initial scheduling. In this study, a dynamic causal diagram and constraint-based scheduling method is proposed to calculate the scheduling amount in order to overcome the drawbacks. Probability which is used to calculate the sufficiency of the continuous data is defined along with the computing method so as to select the most related variables to the gas tank level. Furthermore, a dynamic causal diagram for BFG system is established according to the selected variables, and an objective function based on the difference between BFG generation and consumption is constructed as well as the inequality constraints of the related variables for calculating the scheduling amount. The experiments using real practical data coming from a steel plant in China indicate that the proposed method can effectively improve the scheduling accuracy and reduce the gas diffusion.

Space direction neighborhood preserving embedding-based monitoring and scheduling guidance for blast furnace gas system

Blast furnace gas (BFG) system of steel enterprise generally accompanies with multi-dimension and nonlinear features. Its a hard assignment for energy scheduling operators to make real-time scheduling decision when monitoring such system. In this study, a novel dimensionality reduction method named Space Direction Neighborhood Preserving Embedding (SDNPE) is proposed for the BFG system monitoring and scheduling units determination. To maintain the system dynamic characteristic in the low dimension space, such method constructs a neighborhood graph that searches for nearest neighbors with respect to both the neighbors in spatial scales and fluctuation tendency of the gas flow data. Then, for the BFG system monitoring and scheduling units determination, Hotellings T2 chart and score chart are constructed upon the SDNPE model. Experiments with real-time data of an iron enterprise in China demonstrated the effectiveness of the proposed method.