基于动态阈值神经网络模型的Flow Shop排序研究

根据神经元的动作特征，提出了一种基于动态阈值的神经网络模型，用于求解Flow Shop排序问题，研究表明，这种模型能简化网络运行的中间过程，修正二值输出函数的性能，模型复杂性的降低使收敛速度和有效性得到了较好的改善，模型具有的模拟退火效果使系统跳出局部最优而收敛于全局最优的可能性增大。

Scheduling Flow Shop Based on ANN Model with Dynamic Threshold

According to the characters of the neuron activities,an ANN model is put forward based on dynamic threshold to schedule flow shop. It is shown in the study that the proposed model can simplify the process of the network operation,correct the performance of 2\|values output functions. Due to the decrease of complexity, the convergence speed and the effectiveness of the method are greatly improved. The model can also make the algorithm to converge in the global optima point more possibility because of the effect of simulated annealing.