Papers by Keyword: Differential Evolution (DE)

Abstract: Emission dispatching is conducted to calculate the lowest amount of emission while generating satisfying output to the load demand. The utilities are restricted by emission regulation that limits the emission level to a certain amount. This paper proposes emission dispatch with multiple fuel option (EDMFO) to determine the optimal emission level. The EDMFO allows the operators to select different type of fuel according to the generation level and requirement. The emission dispatch problem is optimized by using Differential Evolution Immunized Ant Colony Optimization (DEIANT) technique. Validation process is conducted by comparing DEIANT with several optimization approaches including ACO and EP. The comparison took places in IEE 57-Bus RTS. Results indicate that DEIANT is superior in terms of calculating the lowest emission level, lower operating cost and the best selection of fuel according to the generation requirement.

Abstract: The micro-vibration mechanism and fault characteristics of micro-gears was described and the faults were classified with no fault, gear crack, gear face wear, tooth face attrition, tooth face crack. The wavelet neural network was proposed and optimized with differential evolution algorithm. The test was taken with the diagnosis information acquired with vibration experiment and designed as training samples which was normalized for wavelet neural networks, the simulation was taken under MATLAB and the simulation result shows the new algorithm with convergence quality and higher diagnosis precision.

Abstract: Tool wear monitor is one of the critical issue in industry, accurate prediction of tool life can guarantee surface quality, the new method of tool wear stated recognition was proposed with application of wavelet neural networks (WNN). An application of differential evolution (DE) algorithm was introduced to training artificial neural networks. The three layer wavelet neural network was constructed and optimized with differential evolution algorithm. Cutting force and cutting noise was monitored and the signal was processed as training samples for wavelet neural networks. Simulation with BPNN, WNN and DEWNN to show the new method can avoid normal neural networks with convergence quality and enhance learning speed, also the diagnosis precision and efficiency was improved.

Abstract: With the disadvantages of volatility, intermittent and randomness of wind power, a research on constructing a fairly accurate prediction model is imperative to improve the quality of power system. Considering the optimization ability of heuristic algorithm and the regression ability of support vector machine, a HA-SVM model is constructed.Case study shows that, compared with other heuristic algorithms, the search efficiency and speed of differential evolution are good, and the prediction accuracy of the model is high.

Abstract: A prediction algorithm for traffic flow prediction of BP neural based on Differential Evolution(DE) is proposed to overcome the problems such as long computing time and easy to fall into local minimum by combing DE and neural network . In the algorithm, DE is used to optimize the thresholds and weights of BP neural network, and the BP neural network is used to search for the optimal solution. The efficiency of the proposed prediction method is tested by the simulation of real traffic flow. The simulation results show that the proposed method has higher precision compared with the traditional BP neural network，so prove it is feasible and effective in the practical prediction of traffic flow.

Abstract: Maintaining gap between Electrode and workpiece in Electrical Discharge Machining (EDM) is very important since the capability of control system to keep the gap will improve the performance of this machine. Therefore to maintain the gap, a Proportional Integral Derivative (PID) controller is designed and applied to EDM servo actuator system. The objective of this work is to obtain a stable, robust and controlled system by tuning the PID controller using Differential Evolution (DE) algorithm. The controller for EDM die sinking is verified by simulation of the control system using MATLAB/Simulink program. Simulation results verify the capabilities and effectiveness of the DE algorithm to search the best configuration of PID parameters controller to control the electrode position.

Abstract: Due to shortcoming of traditional image matching for computing the fitness for every pixel in the searching space, a new bat algorithm with mutation (BAM) is proposed to solve image matching problem, and a modification is applied to mutate between bats during the process of the new solutions updating. This new approach can accelerate the global convergence speed while preserving the strong robustness of the basic BA. The realization procedure for this improved meta-heuristic approach BAM is also presented. To prove the performance of this proposed meta-heuristic method, BAM is compared with BA and other population-based optimization methods, DE and SGA. The experiment shows that the proposed approach is more effective and feasible in image matching than the other model.

Abstract: The optimization design about hybrid discrete variables is very significant but also difficult in engineering, mathematics for programming and operational research. Aimed at shortages of existing optimum methods, in this paper, according to the search mechanism of differential evolution algorithm (DEA), a new differential evolution algorithm is proposed to complex optimization problem with hybrid discrete variables The dynamic penalty function was constructed. DEA algorithm program with hybrid discrete variables is developed. The computing examples of mechanical optimization design show that this algorithm has no special requirements on the characteristics of optimal designing problems, it has a fairly good universal adaptability and a reliable operation of program with a strong ability of overall convergence and high efficiency.

Abstract: A differential evolution algorithm based job scheduling method is presented, whose optimization target is production cost. The cost optimization model of hybrid flow-shop is thereby constructed through considering production cost as a factor in scheduling problem of hybrid flow-shop. In the implementation process of the method, DE is used to take global optimization and find which machine the jobs should be assigned on at each stage, which is also called the process route of the job; then the local assignment rules are used to determine the job’s starting time and processing sequence at each stage. With converting time-based scheduling results to fitness function which is comprehensively considering the processing cost, waiting costs, and the products storage costs, the processing cost is taken as the optimization objective. The numerical results show the effectiveness of the algorithm after comparing between multi-group programs.

Abstract: An Improved Differential evolution (IDE) is proposed in this paper. It has some new features: 1) using multi-parent search strategy and stochastic ranking strategy to maintain the diversity of the population; 2) a novel convex mutation to accelerate the convergence rate of the classical DE algorithm.; The algorithm of this paper is tested on 13 benchmark optimization problems with linear or/and nonlinear constraints and compared with other evolutionary algorithms. The experimental results demonstrate that the performance of IDE outperforms DE in terms of the quality of the final solution and the stability.