智能制造系统调度的一种随机神经网络求解法

A Method of Stochastic Neural Network for Solving IMS Scheduling

首先提出了智能调度系统的概念，即智能制造系统智能调度应从整体指标出发，综合研究集零件加工智能调度和产品装配智能规划于一体的调度技术；然后，给出了智能制造系统智能调度的神经网络模型的能量函数，并讨论了一种改进的模拟退火过程（ＩＡＰ）来求解最优解的方法；最后给出了仿真结果。仿真结果表明，这种改进的模拟退火算法既具有较快的求解速度，又可以逃逸局部最优，而求得问题的全局最优解。由于这种调度方法是柔性的，因此，它可以应用于各种调度系统中。

Intelligent scheduling system is the most important part of intelligent manufacturing system (IMS). Authors firstly propose the conception of intelligent scheduling for the intelligent manufacture system which is an integration of scheduling for flexible manufacture system (FMS) and for flexible assembly system (FAS). Intelligent scheduling for IMS is flexible and hightly intelligent. A formulation for the IMS scheduling problem and a novel algorithm, improved simulated annealing (IAP), which is the optimum solution can be obtained, were intrduced. The simulated result which is given in final section shows that IAP can dislodge a state from a local minimum and guides it to the global minimum. This scheduling approach is flexible. It can be used not only in IMS in future but also other scheduling situation.