一种结合IPSO-BP神经网络的备件供应模型研究

Spare Parts Supply Model Based on IPSO-BP Neural Network is Studied

备件供应是制造业服务价值链协同中的重要组成,也是企业制定销售计划和生产计划的重要依据.本文将备件供应过程中的备件消耗考虑在内,以最小化总成本为目标,以订货时的备件需求量为核心参数,提出一种基于神经网络的备件供应需求预测模型.在现有标准粒子群算法的基础上,通过将惯性权重的改进、环境检测策略和自适应最优解跳跃策略结合,提出一种改进的粒子群算法(IPSO,Improved Particle Swarm Optimization).并通过改进的粒子群算法对BP(Back Propagation)神经网络进行优化.最后通过IPSO-BP神经网络模型对备件供应模型中的备件需求量进行预测,实验结果表明,相比其他的神经网络模型,IPSO-BP神经网络模型的预测稳定性和精准度等性能有显著提高.

Spare parts supply is an important component of manufacturing service value chain coordination,and also an important basis for enterprises to make sales plan and production plan.Taking the consumption of spare parts in the process of spare parts supply into consideration,aiming at minimizing the total cost and taking the spare parts demand at the time of ordering as the core parameter,this paper proposes a spare parts supply demand forecasting model based on neural network.An Improved Particle Swarm Optimization algorithm(IPSO,Improved Particle Swarm Optimization)is proposed based on the existing standard Particle Swarm Optimization(PSO),which combines the Improved inertia weight,environment detection strategy and self-adaptive jumping strategy.And optimize the BP(Back Propagation)neural network through improved particle swarm optimization.Finally,the demand for spare parts in the spare parts supply model was predicted by IPSO-BP neural network model,the experiment result shows that,compared to other neural network models,the prediction stability and accuracy of IPSO-BP neural network model are significantly improved.