一种“工期固定,资源均衡”的神经网络模型设计

A DESIGN ON A NEURAL NETWORK FOR RESOURCE BALANCE OPTIMIZATION

使用网络计划进行工程进度的控制 ,通常需要经过优化才能达到好的效果。本文根据缩小方差法的思想 ,提出了把网络图直接映射为神经网络的思想 ,并设计了一个可用于工期固定 ,单一资源均衡的神经网络 ;最后通过实例表明了该网络具有理想的优化性能。

Network planning can be using for progress control,but it needs further optimization.According to the reduced variance method,a new method mapping the network diagram into neural network is proposed.Meanwhile,a neural network specially designed for “fixed term,single resource balance” is established and the result tested by a test shows its satisfactory performance.