论文部分内容阅读
本文介绍了用神经网络求解FMS中有约束的资源调度问题的方法。有约束的资源调度问题首先被分解成一系列多维背包模型并且为背包模型建立了一个等价的Hopfield神经网络,然后通过扩展Hopfield网络,给出了一种求解有约束的资源调度问题的方法。这种方法可以避免通常神经网络所具有的不稳定性和容易陷入局部极小点的缺陷。
This paper introduces a neural network to solve the constrained resource scheduling problem in FMS. The constrained resource scheduling problem is first decomposed into a series of multidimensional backpacking models and an equivalent Hopfield neural network is established for the backpacking model. Then, by extending the Hopfield network, a solution to the constrained resource scheduling problem is given. This method can avoid the usual neural network instability and easy to fall into local minimum defects.