Enhancing Supply Chain Agility and Sustainability through Machine Learning: Optimization Techniques for Logistics and Inventory Management
Abstract
:1. Introduction
1.1. State of the Art
1.2. Problem Statement
- Developing advanced machine learning models that can effectively process and analyze both structured and unstructured data from diverse supply chain activities.
- Evaluating the ability of these models to dynamically adapt to changing conditions and accurately predict supply chain needs, from demand forecasting to resource allocation.
- Comparing the performance of these models against traditional supply chain management approaches to quantify improvements in efficiency, cost reduction, and decision-making accuracy.
2. Materials and Methods
2.1. Overview of Algorithms
- Linear Regression: This is the simplest form of regression, used for predicting a continuous dependent variable based on one or more independent variables [22]. It assumes a linear relationship between input (predictors) and output (response).
- Ridge Regression: This method extends linear regression by adding a regularization penalty to the loss function [23]. This penalty shrinks the coefficients of correlated predictors and is particularly useful in scenarios where the prediction model suffers from high multicollinearity or when the number of predictors exceeds the number of observations.
- Lasso Regression: Lasso, or Least Absolute Shrinkage and Selection Operator, introduces a regularization term that not only helps in reducing overfitting but also performs feature selection [24].
- Elastic Net Regression: Combining the penalties of ridge and lasso regression, elastic net is particularly useful when dealing with highly correlated data [25]. It can reduce the variability of coefficients estimated by ordinary least squares and is robust against overfitting in a model with many predictors.
- Gradient Boosted Trees: This ensemble technique builds models in stages, like other boosting methods, and generalizes them by allowing optimization of an arbitrary differentiable loss function [26].
2.1.1. Clustering Algorithms
- K-means Clustering: Known for its efficiency and simplicity, K-means clustering will be used to segment large datasets based on similarities within the data, aiding in operational optimizations such as inventory categorization and risk management [27].
- DBSCAN (Density-Based Spatial Clustering of Applications with Noise): This algorithm is particularly useful for identifying outliers and handling irregularly shaped clusters [28]. In supply chain management, DBSCAN will be applied to detect and analyze atypical patterns or anomalies in logistical data, enhancing risk monitoring capabilities.
2.1.2. Neural Networks
- Convolutional Neural Networks (CNNs): These are utilized for their superior ability to process grid-like data, including images and spatial structures [29]. In logistics optimization, CNNs will analyze route maps and traffic patterns to recommend optimal transportation routes and schedules.
- Recurrent Neural Networks (RNNs) with Long Short-Term Memory (LSTM) units: Ideal for time-series prediction, LSTMs will be deployed to forecast demand and manage inventory levels by learning from historical sales data and external factors like market trends and seasonal fluctuations [30].
- Feedforward Neural Networks with Attention Mechanisms: These networks will be examined for their ability to enhance model interpretability and handle complex multivariate time series data, crucial for dynamic risk assessment and decision-making processes in supply chain management [31].
2.2. Data Sources and Collection
- Transportation Data: Sourced from a global logistics provider, this dataset includes comprehensive records of 500,000 shipment transactions, detailing pickup and delivery locations, shipment dates, weights, transportation modes, and carriers. This structured dataset is pivotal for modeling logistics optimization tasks such as route planning and freight management.
- Inventory Data: Obtained from an omni-channel retailer, this dataset encompasses two years of item-level sales, demand fulfillment, and replenishment transactions covering 10,000 Stock Keeping Units (SKUs). It provides a granular view of inventory dynamics necessary for demand forecasting and stock level optimization.
- External Data: Includes public traffic and road closure alerts, which consist of about 2 million alerts per month, and extensive social media discussions and news archives related to supply chain issues, totaling approximately 10 million documents. These unstructured data enrich the models with external contextual factors affecting supply chain performance.
2.3. Data Processing Steps
- Data Cleaning: Identifying and correcting inaccuracies or inconsistencies in the data, such as missing values or duplicate records. This step ensures the quality and reliability of the models’ inputs.
- Data Transformation: Converting raw data into a format suitable for analysis. This may involve normalizing data scales, encoding categorical variables, or generating datetimes features from timestamps.
- Feature Engineering: Creating new variables by combining or transforming existing features to enhance model performance. Techniques such as PCA (Principal Component Analysis) for dimensionality reduction or creating interaction terms between features might be applied, especially in handling high-dimensional data like inventory and transportation records.
- Integration: Combining different data sources into a unified dataset. This often requires aligning data on common identifiers, reconciling discrepancies between related datasets, and ensuring synchronized time frames across datasets.
2.4. Performance Evaluation
2.4.1. Statistical Testing
- Hypothesis Testing: Techniques such as t-tests or ANOVA will be used to statistically validate the improvements attributed to the machine learning models [36]. These tests will help confirm that the observed enhancements in supply chain performance metrics are significant and not due to random variation.
- Confidence Interval Analysis: By calculating confidence intervals around performance metrics, the study will quantify the uncertainty in the estimates provided by the models, offering insights into their reliability and the range of expected outcomes [37].
2.4.2. Business Performance Metrics
- Inventory Cost Reduction: By optimizing reorder points and stock levels, the models aim to minimize holding costs and reduce the likelihood of stockouts, directly impacting the bottom line.
- Delivery Efficiency: Models that improve routing and scheduling are assessed based on their ability to reduce delivery times and enhance on-time delivery rates, crucial for customer satisfaction and operational efficiency.
- Supply Chain Resilience: Enhanced predictive capabilities can lead to better anticipation of supply chain disruptions and quicker response times, thereby improving the overall resilience of the supply chain network.
3. Results
3.1. Application of Machine Learning Models
3.1.1. Transportation Optimization
3.1.2. Demand Forecasting and Inventory Optimization
3.2. Statistical Evaluation of Model Performances
Comparison of Models
3.3. Risk Monitoring and Production Scheduling
Production Efficiency
3.4. Comparative Analysis across Applications
4. Discussion
4.1. Practical Implications
4.2. Limitations and Challenges
4.3. Future Research Directions
5. Conclusions
Author Contributions
Funding
Data Availability Statement
Conflicts of Interest
References
- Curcio, D.; Longo, F. Inventory and internal logistics management as critical factors affecting the supply chain performances. Int. J. Simul. Process Model. 2009, 5, 278–288. [Google Scholar] [CrossRef]
- Oluwaseyi, J.A.; Onifade, M.K.; Odeyinka, O.F. Evaluation of the role of inventory management in logistics chain of an organisation. LOGI—Sci. J. Transp. Logist. 2017, 8, 1–11. [Google Scholar] [CrossRef]
- Sharma, N.; Singhi, R. Logistics and supply chain management quality improvement of supply chain process through vendor managed inventory: A QFD approach. J. Supply Chain. Manag. Syst. 2018, 7, 23–33. [Google Scholar]
- Gölzer, P.; Fritzsche, A. Data-driven operations management: Organisational implications of the digital transformation in industrial practice. Prod. Plan. Control 2017, 28, 1332–1343. [Google Scholar] [CrossRef]
- Mahraz, M.I.; Benabbou, L.; Berrado, A. Machine learning in supply chain management: A systematic literature review. Int. J. Supply Oper. Manag. 2022, 9, 398–416. [Google Scholar]
- Karim, M.R.; Beyan, O.; Zappa, A.; Costa, I.G.; Rebholz-Schuhmann, D.; Cochez, M.; Decker, S. Deep learning-based clustering approaches for bioinformatics. Brief. Bioinform. 2021, 22, 393–415. [Google Scholar] [CrossRef]
- Caton, S.; Malisetty, S.; Haas, C. Impact of imputation strategies on fairness in machine learning. J. Artif. Intell. Res. 2022, 74, 1011–1035. [Google Scholar] [CrossRef]
- Malisetty, S.; Ali, H.H.; Rastegari, E.; Siu, K.C. An Innovative Comparative Analysis Approach for the Assessment of Laparoscopic Surgical Skills. Surgeries 2023, 4, 46–57. [Google Scholar] [CrossRef]
- Thelagathoti, R.K.; Malisetty, S.; Ali, H.H. Analyzing Walking and Driving Behavior Across Different Age Groups Using Population Analysis and Correlation Networks. In Proceedings of the 2022 5th International Conference on Communications, Signal Processing, and their Applications (ICCSPA), Cairo, Egypt, 27–29 December 2022; IEEE: Piscataway, NJ, USA, 2022; pp. 1–6. [Google Scholar]
- Abolghasemi, M.; Beh, E.; Tarr, G.; Gerlach, R. Demand forecasting in supply chain: The impact of demand volatility in the presence of promotion. Comput. Ind. Eng. 2020, 142, 106380. [Google Scholar] [CrossRef]
- Wang, J.; Ma, Y.; Zhang, L.; Gao, R.X.; Wu, D. Deep learning for smart manufacturing: Methods and applications. J. Manuf. Syst. 2018, 48, 144–156. [Google Scholar] [CrossRef]
- Alam, M.; Alam, M.S.; Roman, M.; Tufail, M.; Khan, M.U.; Khan, M.T. Real-time machine-learning based crop/weed detection and classification for variable-rate spraying in precision agriculture. In Proceedings of the 2020 7th International Conference on Electrical and Electronics Engineering (ICEEE), Antalya, Turkey, 14–16 April 2020; IEEE: Piscataway, NJ, USA, 2020; pp. 273–280. [Google Scholar]
- Huang, G.B.; Zhu, Q.Y.; Siew, C.K. Real-time learning capability of neural networks. IEEE Trans. Neural Netw. 2006, 17, 863–878. [Google Scholar] [CrossRef]
- Morariu, C.; Morariu, O.; Răileanu, S.; Borangiu, T. Machine learning for predictive scheduling and resource allocation in large scale manufacturing systems. Comput. Ind. 2020, 120, 103244. [Google Scholar] [CrossRef]
- Usuga Cadavid, J.P.; Lamouri, S.; Grabot, B.; Pellerin, R.; Fortin, A. Machine learning applied in production planning and control: A state-of-the-art in the era of industry 4.0. J. Intell. Manuf. 2020, 31, 1531–1558. [Google Scholar] [CrossRef]
- Jayaprakash, S.; Nagarajan, M.D.; Prado, R.P.D.; Subramanian, S.; Divakarachari, P.B. A systematic review of energy management strategies for resource allocation in the cloud: Clustering, optimization and machine learning. Energies 2021, 14, 5322. [Google Scholar] [CrossRef]
- Yang, M.; Lim, M.K.; Qu, Y.; Ni, D.; Xiao, Z. Supply chain risk management with machine learning technology: A literature review and future research directions. Comput. Ind. Eng. 2023, 175, 108859. [Google Scholar] [CrossRef]
- Theodorou, E.; Spiliotis, E.; Assimakopoulos, V. Optimizing inventory control through a data-driven and model-independent framework. EURO J. Transp. Logist. 2023, 12, 100103. [Google Scholar] [CrossRef]
- Bertolini, M.; Mezzogori, D.; Neroni, M.; Zammori, F. Machine Learning for industrial applications: A comprehensive literature review. Expert Syst. Appl. 2021, 175, 114820. [Google Scholar] [CrossRef]
- Odimarha, A.C.; Ayodeji, S.A.; Abaku, E.A. Machine learning’s influence on supply chain and logistics optimization in the oil and gas sector: A comprehensive analysis. Comput. Sci. IT Res. J. 2024, 5, 725–740. [Google Scholar] [CrossRef]
- Elahi, M.; Afolaranmi, S.O.; Martinez Lastra, J.L.; Perez Garcia, J.A. A comprehensive literature review of the applications of AI techniques through the lifecycle of industrial equipment. Discov. Artif. Intell. 2023, 3, 43. [Google Scholar] [CrossRef]
- Uyanık, G.K.; Güler, N. A study on multiple linear regression analysis. Procedia Soc. Behav. Sci. 2013, 106, 234–240. [Google Scholar] [CrossRef]
- Saunders, C.; Gammerman, A.; Vovk, V. Ridge regression learning algorithm in dual variables. In Proceedings of the 15th International Conference on Machine Learning, ICML, Madison, WI, USA, 24–27 July 1998. [Google Scholar]
- Ranstam, J.; Cook, J.A. LASSO regression. J. Br. Surg. 2018, 105, 1348. [Google Scholar] [CrossRef]
- Hans, C. Elastic net regression modeling with the orthant normal prior. J. Am. Stat. Assoc. 2011, 106, 1383–1393. [Google Scholar] [CrossRef]
- Natekin, A.; Knoll, A. Gradient boosting machines, a tutorial. Front. Neurorobot. 2013, 7, 21. [Google Scholar] [CrossRef]
- Sinaga, K.P.; Yang, M.S. Unsupervised K-means clustering algorithm. IEEE Access 2020, 8, 80716–80727. [Google Scholar] [CrossRef]
- Khan, K.; Rehman, S.U.; Aziz, K.; Fong, S.; Sarasvady, S. DBSCAN: Past, present and future. In Proceedings of the Fifth International Conference on the Applications of Digital Information and Web Technologies (ICADIWT 2014), Chennai, India, 17–19 February 2014; IEEE: Piscataway, NJ, USA, 2014; pp. 232–238. [Google Scholar]
- Li, Z.; Liu, F.; Yang, W.; Peng, S.; Zhou, J. A survey of convolutional neural networks: Analysis, applications, and prospects. IEEE Trans. Neural Netw. Learn. Syst. 2021, 33, 6999–7019. [Google Scholar] [CrossRef]
- Medsker, L.R.; Jain, L. Recurrent neural networks. Des. Appl. 2001, 5, 2. [Google Scholar]
- Svozil, D.; Kvasnicka, V.; Pospichal, J. Introduction to multi-layer feed-forward neural networks. Chemom. Intell. Lab. Syst. 1997, 39, 43–62. [Google Scholar] [CrossRef]
- Chicco, D.; Warrens, M.J.; Jurman, G. The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. PeerJ Comput. Sci. 2021, 7, e623. [Google Scholar] [CrossRef]
- Tatachar, A.V. Comparative assessment of regression models based on model evaluation metrics. Int. Res. J. Eng. Technol. (IRJET) 2021, 8, 853–860. [Google Scholar]
- Shahapure, K.R.; Nicholas, C. Cluster quality analysis using silhouette score. In Proceedings of the 2020 IEEE 7th International Conference on Data Science and Advanced Analytics (DSAA), Sydney, Australia, 6–9 October 2020; IEEE: Piscataway, NJ, USA, 2020; pp. 747–748. [Google Scholar]
- Shutaywi, M.; Kachouie, N.N. Silhouette analysis for performance evaluation in machine learning with applications to clustering. Entropy 2021, 23, 759. [Google Scholar] [CrossRef]
- Shaw, R.G.; Mitchell-Olds, T. ANOVA for unbalanced data: An overview. Ecology 1993, 74, 1638–1645. [Google Scholar] [CrossRef]
- Figueiras, A.; Domenech-Massons, J.M.; Cadarso, C. Regression models: Calculating the confidence interval of effects in the presence of interactions. Stat. Med. 1998, 17, 2099–2105. [Google Scholar] [CrossRef]
- Akbari, M.; Do, T.N.A. A systematic review of machine learning in logistics and supply chain management: Current trends and future directions. Benchmarking Int. J. 2021, 28, 2977–3005. [Google Scholar] [CrossRef]
- Suwignjo, P.; Panjaitan, L.; Baihaqy, A.; Rusdiansyah, A. Predictive analytics to improve inventory performance: A case study of an FMCG Company. Oper. Supply Chain. Manag. Int. J. 2023, 16, 293–310. [Google Scholar] [CrossRef]
Data Source | Size/Volume |
---|---|
Logistics Provider Transportation Data | 500,000 shipment records |
Public Traffic and Road Closure Alerts | 2 million alerts per month |
News Archives and Social Media Discussions | 10 million documents |
Omni-channel Retailer Item Sales and Transactions | 2 years of data, 10,000 SKUs |
Item Attributes, Promotions, Weather Data | 100,000 documents |
Manufacturer Production Records | 5 years of data, 20 plants |
Financial Reports, Press Releases, Economic Indicators | 50,000 documents |
Geo-political Events | 10,000 events |
Model | Metric | Value |
---|---|---|
CNN Transportation Optimization | On-time Delivery Percentage (Model) | 98% |
On-time Delivery Percentage (Historical) | 94% | |
t-statistic | 23.54 | |
p-value | <0.001 | |
Cost Savings from Reduced Rerouting | 6% per shipment | |
LSTM Demand Forecaster | Mean Absolute Percentage Error (LSTM) | 2.3% |
Mean Absolute Percentage Error (ARIMA) | 3.7% | |
Mean Absolute Percentage Error (ETS) | 4.1% | |
F-statistic | 1245.61 | |
p-value | <0.001 |
Model | Metric | Value |
---|---|---|
Demand Forecasting Accuracy | Mean Absolute Error (Gradient Boosted Trees) | 1105 units |
Mean Absolute Error (Linear Regression) | 1325 units | |
t-statistic | 123.56 | |
p-value | <0.001 | |
Inventory Optimization | Inventory Level Reduction | 5–10% |
Service Level | 99% | |
Adaptability | Dynamic |
Model | Metric | Traditional (ARIMA, ETS) | ML (LSTM, CNN) |
---|---|---|---|
Demand Forecasting Accuracy | Mean Absolute Percentage Error (MAPE) | 4.1% (ETS), 3.7% (ARIMA) | 2.3% (LSTM) |
Inventory Optimization | Reduction in Overstock/Stockouts | 5% | 10% |
Delivery Optimization | On-time Delivery Rate | 94% | 98% |
Cost Savings in Logistics | Cost Reduction per Shipment | 2% | 6% |
Study | Model | Pros | Cons |
---|---|---|---|
Current Study | LSTM, CNN | High accuracy, adaptability, reduced costs | Data dependency, implementation complexity |
[38] | ARIMA, ETS | Simplicity, well-understood | Limited to linear relationships, less accurate |
[39] | Regression | Easy to implement, interpretable | High error rates, not suitable for non-linear data |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Pasupuleti, V.; Thuraka, B.; Kodete, C.S.; Malisetty, S. Enhancing Supply Chain Agility and Sustainability through Machine Learning: Optimization Techniques for Logistics and Inventory Management. Logistics 2024, 8, 73. https://doi.org/10.3390/logistics8030073
Pasupuleti V, Thuraka B, Kodete CS, Malisetty S. Enhancing Supply Chain Agility and Sustainability through Machine Learning: Optimization Techniques for Logistics and Inventory Management. Logistics. 2024; 8(3):73. https://doi.org/10.3390/logistics8030073
Chicago/Turabian StylePasupuleti, Vikram, Bharadwaj Thuraka, Chandra Shikhi Kodete, and Saiteja Malisetty. 2024. "Enhancing Supply Chain Agility and Sustainability through Machine Learning: Optimization Techniques for Logistics and Inventory Management" Logistics 8, no. 3: 73. https://doi.org/10.3390/logistics8030073