| 黄育鹏， 何雪明， 卢立新， 林自东.改进粒子群算法在库存预测中的应用[J].轻工机械,2022,40(2):103-108 |
| 改进粒子群算法在库存预测中的应用 |
| Application of Improved Particle Swarm Optimization in Inventory Forecasting |
| |
| DOI：10.3969/j.issn.1005 2895.2022.02.017 |
| 中文关键词: 立体仓库 库存预测 粒子群算法 指数平滑法 异步变化学习因子 |
| 英文关键词:stereoscopic warehouse inventory forecast PSO（Particle Swarm Optimization） exponential smoothing asynchronous change learning factor |
| 基金项目:国家自然科学基金资助项目(51275210)；国家自然科学基金资助项目(51975251)；省产学研联创项目(1078081606192480)；江苏省食品先进制造装备技术重点实验室自主研究课题资助项目（FMZ2018Y2）；“六大人才高峰”(1076010241131350)。[ZK）] |
|
| 摘要点击次数: 180 |
| 全文下载次数: 167 |
| 中文摘要: |
| 针对自动化立体仓库库存预测结果存在不准确和时间滞后等问题，课题组提出一种基于改进粒子群算法并结合指数平滑法来构建库存预测模型。分析传统粒子群算法和指数平滑法的原理以及缺点，通过引入附加变量、非线性动态调整惯性权重以及异步变化学习因子的方式，提出一种改进的粒子群算法；并采用4种标准测试函数来验证算法的寻优能力；最后将改进后的算法与平滑指数算法相结合构建预测模型，以某公司生产的导流板实际库存数据为例进行仿真实验，并与常用的几种预测模型进行验证对比。结果表明改进的粒子群算法预测模型的精度更高。该模型能够解决传统预测模型精度不高、适用情况单一等问题，提高企业的库存利用率。 |
| 英文摘要: |
| Aiming at the problems of the inaccuracy and time lag of inventory forcast results in automated stereoscopic warehouse, an inventory forcast model based on improved particle swarm optimization and exponential smoothing method was proposed. Firstly, the principles and disadvantages of traditional particle swarm optimization algorithm and exponential smoothing method were analyzed. Secondly, an improved particle swarm optimization algorithm was proposed by introducing additional variables, nonlinear dynamic adjustment of inertia weight and asynchronous change learning factors. Four standard test functions were used to verify the optimization ability of the algorithm. Finally, the improved algorithm was combined with the exponential smoothing algorithm to build the prediction model. The simulation experiment was carried out by taking the actual inventory data of a deflector produced by a company as an example, and the built prediction model was compared with several commonly used prediction models for validation. The results show that the accuracy of the improved particle swarm optimization prediction model is higher. This model can solve the problems of insufficient accuracy and single application of the traditional prediction model, and improve the inventory utilization rate of enterprises. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
