一、什么是供应链?
概况的说:
流动(物体/金融/信息)———正反向物流(购买/退货)
供应链:物料供给,采购,生产设备,物流,市场
提高效率———作用:增加价值链,利益最大化
目的: 客户服务
二、AI 现目前最流行在SCM方面
Marketing:
1.sales方面:
ANN:神经网络,预测销售额
Agent based system(ABSs):实时 销售管理
Genetic algorithm:线上促销
2.pricing方面:
ANN:价格模型
AI科技:多样生产和服务
3.segmentation方面:(eg:拆分客户--通过consumer decision behavior;拆分市场)
cluster聚类模型:k-means+ANN
fuzzy模型
4. facilitate the selection of contacts方面:
support vector data 用支撑向量机描述
5.directed marketing(直销市场)
6.product life-cyle management(产品生命周期)-先冷链物流比较火
Logistics-物体的流通和储存
流通的承载物:车/container集装箱
系统:系统自动化,
工作流通中:和作业相关的RFID(radio-frequency identification)-增强工作流反映速度
Production:
1.assenly line相关的
2.production monitoring
3.production forcasting方面
neural decision tree(NDT)/ANN
Gamma classifier +feedforward neural network:时间序列的生产预测
4.production system方面
self-learning production ramp-up system(生产更新系统)
decentrailized data analysis(去中心数据分析--用于数据提取)
5.production planning方面
AI-based classification methodology(分类):ANN/GA/data mining
6.manufacturing system制造系统
fuzzy logic(FL)
Gaussian model
7.manufacturing decision 制造决策
CBR(case-based reasoning)
RBR(rule-based reasoning)
8.quality monitoring 质量监控
Supply chain:
1.demand forcasting(需求预测)
ANN/fuzzy inference system(对信息不完全的进行模糊预测)
基于客户细分-
支持向量机(SVM)-的时间序列分类(time series)
artificial neural network model(人工神经网络)-预测需求
2.灾害管理(humanitarian facility location--disaster)
物资分配-和network/location 有关
3.supplier selection供应商的选择方面
fuzzy-bayesian supplier selection
neuro-fuzzy supplier selection
4.supplychain planning
ABSs-实时销售+MASs---simulation
5.competitive oligopolistic market(寡头竞争市场)
supply chain network 最优化
6.risk monitoring system(RMS)
基于物联网
三、模型总结
1.在生产中:ANNs,GA,ABSs,FL常用
ANNs:信息处理技术,基于回归,大量数据中找到模型(销售预测,营销DSS,定价,客户细分,生产预测,供应商选择,需求预测)
FL(fuzzy logic):做决策,判断的
ABSs(agent-based model)-MAs(a network of agents):用于分布式供应链规划,系统设计模拟,供应链风险
GAs:多目标优化,supplier/cooperator 的选择
Data mining :用数据洞悉决策:控制/监控仓库,食品供应链,供应链可持续发展
CBR:基于认知心理观念的技术(case-based),通过多个问题解决一个问题。风险管理,绩效评估,供应链谈判,agile SCM,
SVMs:数据进行分类,线性分类器。供应商选择,需求预测,时间序列分类
K-means:聚类;heuristic:启发式,stomatic simulation:随机性模拟;bayesian:需求预测;decision tree:决策树;Gaussian model:高斯
markov decision process:做决策
2.预测模型,判断决策模型,模拟实时模型 常用 。
3.Marketing
1.Artificial neural networks:(4) 2. Genetic algorithm (4)
3. FL/modelling (3) 4. Agent-based/multi-agent systems (2)
5. Swarm intelligence (1) 6. Simulated annealing (1)
7.Association rule (1) 8.Tree-based models (1)
9.Support vector machines (1) 10.General forms of AI (1)
11.k-means clustering (1) 12. Hill climbing (1)
4.Logistics
1.Artificial neural networks (1) 2.Agent-based/multi-agent systems (1)
3.Data mining (1) 4.Simulated annealing (1)
5.Automated planning (1) 6. Robot programming (1)
7. General forms of AI (1) 8. Heuristics (1)
5.Production
1.Artificial neural networks (8) 2.FL/modelling (5)
3. Case-based reasoning (4) 4. Genetic algorithm (3)
5.Agent-based/multi-agent systems (2) 6. Data mining (2)
7.Decision trees (2) 8.General forms of AI (1)
9. Gaussian (1) 10. Rule-based reasoning (1)
11. Automated planning (1) 12.Swarm intelligence
6.supplychain
1.Artificial neural networks (5) 2.FL/modelling (4)
3. Agent-based/multi-agent systems (4) 4. General forms of AI (4)
5.Physarum model (1) 6. Bayesian networks (1)
7.Swarm intelligence (1) 8.Data mining (1)
9.Support vector machines (1) 10. Stochastic simulation (1)
四、research Gap!
1.disaster management--用ABSs
2.supplychain risks -用ABSs
3.real-time pricing(实时定价):关注非中国领域
4.researse auctioning(反向拍卖卖)-定价
5.supplychian+blockchain /IoT
6.cost optimization:facility location+AI
7.研究方向:无人机,重型车辆,邮政服务,旅游业,汽车+AI
参考文献:Toorajipour, R., Sohrabpour, V., Nazarpour, A., Oghazi, P., & Fischl, M. (2021). Artificial intelligence in supply chain management: A systematic literature review. Journal of Business Research, 122, 502-517.