E-Commerce Workshop Scheduling Based on Deep Learning and Genetic Algorithm
Peijian Wu, Dar-Li Yang
Abstract
With the gradual rise of customized manufacturing, the connection between e-commerce and intelligence manufacturing system have been deepened, highlighting the importance of intelligent scheduling to both intelligent manufacturing and e-commerce. The key to intelligence manufacturing lies in workshop scheduling. This paper optimizes the genetic algorithm (GA) with deep learning neural network (DLNN) and applies the optimized GA to realize intelligent workshop scheduling. Firstly, the production methods of e-commerce products were analysed, as well as the features of workshop scheduling problem (WSP). On this basis, the authors established a mathematical model of the WSP. Considering the actual needs of the workshop, an integrated scheduling algorithm was designed combining DLNN and GA. The algorithm improves the GA with a DLNN called long short-term memory network (LSTM) and constructs the fitness function in a novel manner. Simulation results show that our algorithm can avoid the local optimal trap that plagues the original GA, and better the global search performance.