Xiu Chun-bo

Adaptive ant colony optimization algorithm

An adaptive ant colony algorithm is proposed to overcome the premature convergence problem in the conventional ant colony algorithm. The adaptive ant colony is composed of three groups of ants: ordinary ants, abnormal ants and random ants. Each ordinary ant searches the path with the high concentration pheromone at the high probability, each abnormal ant searches the path with the high concentration pheromone at the low probability, and each random ant randomly searches the path regardless of the pheromone concentration. Three groups of ants provide a good initial state of pheromone trails together. As the optimization calculation goes on, the number of the abnormal ants and the random ants decreases gradually. In the late optimization stage, all of ants transform to the ordinary ants, which can rapidly concentrate to the optimal paths. Simulation results show that the algorithm has a good optimization performance, and can resolve traveling salesman problem effectively.

Improved super-resolution algorithm of single-frame image based on least square method

In order to improve the quality of single-frame image super-resolution reconstruction, a novel interpolation algorithm based on weighted least squares is proposed. Interpolation pixel can be calculated from different directions by Newton interpolation formula, and get Multiple interpolation results. The interpolation result in the super-solution image can be got by using weighted least squares to fuse all the interpolation results. The weight matrix in the least squares method is determined according to the correlation among the pixels in the original image. Thus, contrasting to the conventional method, it uses more useful information to reconstruct the super-resolution image. Simulation results prove that the method can get the better super-resolution image according to not only visual impression but also the objective evaluation index.

New genetic algorithm improved and its applications

A novel genetic algorithm, named double population genetic algorithm (DPGA), is proposed to improve the performance of the conventional genetic algorithm. An elaborate searching space around the current optimal solution is divided from the original searching space. One small population executes genetic operators to speed up the convergence of the algorithm in the elaborate searching space. And the boundaries of the elaborate searching space are reduced continuously to enhance the searching density during the optimization. Another big population executes genetic operators to ensure the global optimal ability of the algorithm. In this way, the algorithm has global searching ability and fast convergence rate. A lot of simulation results prove that the algorithm can accelerate searching rate, enhance the searching efficiency, and give satisfied results to function optimization problems.

Hysteretic neural network and its application in the prediction of the wind speed series

In order to improve the information processing capabilities of the traditional neural network, and improve the forecast accuracy of the wind speed series, a new hysteretic neural network based on the hysteretic neurons is proposed. The hysteretic neuron is constructed by adding a hysteretic operator into the activation function. The hysteretic characteristic can make the response of the neuron is related to not only the current input information but also the history input information. In this way, the amount of information used in the prediction process is increased, and the prediction ability of the neural network can be improved. In order to avoid the information redundancy or loss, the structure and the training samples are determined according to the autocorrelation function and the partial autocorrelation function of the wind speed series. Simulation results show that the hysteretic neural network can complete the prediction of the wind speed series, and the prediction performance is superior to that of the same type neural network.

Time series predication based on genetic chaotic operators network

Scientifically prediction of some statistical data in practical production can guide mission planning and scheduling, policy-making and emergency treatment. A new dynamic prediction network is proposed to improve the prediction performance of conventional method. The prediction network is composed of many chaotic operators, and its control parameters are optimized by genetic algorithm. The dynamic characteristic of the network can be changed to follow that of the system predicted. The prediction results of actual data, such as passenger traffic, freight traffic, goods volume, and passenger volume, show that the method is valid, and it has good predictive ability and precision.