OPTIMISATION OF LOGISTIC OPERATIONS USING AI

International Journal on Soft Computing, Artificial Intelligence and Applications (IJSCAI),
Vol.13, No.2, May 2024
DOI : 10.5121/ijscai.2024.13201 1
OPTIMISATION OF LOGISTIC OPERATIONS USING AI
Semen Levin
Department of Automated Control Systems, Tomsk State University of Control Systems
and Radio electronics, Tomsk, Russia
ABSTRACT
In today's world, optimising logistical operations becomes a defining factor in the success of retailers. The
complexity and dynamism of modern supply chains require the application of advanced technologies for
accurate demand forecasting and inventory management. Artificial Intelligence (AI) demonstrates immense
potential in addressing these challenges, offering a new level of efficiency and adaptability. This research
focuses on applying AI to analyse a retailer's operational data to optimise its logistical operations.
Utilising analysis models such as SARIMA for time series and random forests for machine learning
enabled precise daily sales forecasts, identification of seasonal peaks, and optimisation of inventory levels
and delivery route planning. Particular attention was paid to the analysis of variable importance,
revealing key factors affecting the efficiency of logistical operations, including historical sales data,
weather conditions, distance to delivery locations, and holiday periods. The investigation results show that
integrating AI into logistics improves forecast accuracy and inventory management efficiency and
contributes to reducing logistical costs and enhancing customer satisfaction by optimising delivery time
and costs. The data-driven and AI-based analytical approach provides a comprehensive view of logistics
management, demonstrating how technological innovations can transform traditional methods and create
a competitive advantage in a dynamic market environment. The research underscores the significance of
adopting data and analytics as the basis for strategic planning and decision-making, reinforcing the idea
that AI serves as a valuable extension of companies' analytical capabilities in the field of logistics.
KEYWORDS
Artificial Intelligence, Logistic Operations, Supply Chain Optimisation, Retail Analytics, Machine
Learning, Demand Forecasting, Inventory Management, Seasonality Analysis, Data Analysis,
Technological Advancements
1. INTRODUCTION
This article presents a research based on data reflecting the dynamics of a retailer over a year,
demonstrating how the application of AI can serve as a foundation for optimising logistic
operations [4]. This data includes information on daily sales, inventory levels, delivery data,
weather conditions, and customer feedback, providing a comprehensive view of the challenges
and needs in inventory management and logistics [5, 6]. Based on this data, the author applied
various methods and tools to create a model to optimise logistic operations. The primary model
used was the random forest machine learning model. It was chosen for its ability to handle many
features and identify nonlinear relationships between variables, which is critically important for
understanding the multifactorial impact on logistics processes. Unlike traditional statistical
methods, the random forest model provides a deeper understanding of complex relationships in
the data.

Vol.13, No.2, May 2024
2
Furthermore, the investigation included an analysis using the random forest model to determine
optimal inventory management strategies and efficient delivery route planning. This approach
reduced logistic costs and improved customer satisfaction by reducing delivery times and
increasing product availability [7-10].
Particular attention was paid to accounting for weather conditions and holidays, which can
significantly impact logistic processes. Integrating these data into the random forest model
enhanced its predictive capability and allowed for more effective responses to external
changes [11, 12].
2. DATA SOURCES AND ANALYTICAL OBJECTIVES
The data utilised in this investigation to optimise logistic operations using Artificial Intelligence
(AI) is derived from an extensive and meticulously curated dataset that tracks a retailer's
operational activities across various dimensions. This dataset encompasses detailed sales records,
delivery logistics, and environmental conditions, each contributing critical insights necessary for
the nuanced application of AI technologies.
Sales Data Detailing
The sales data is categorically segmented across ten distinct product categories, enabling precise
analysis of consumer purchasing patterns and demand fluctuations. The dataset records daily
sales figures for each category, providing a granular view of consumer transactions throughout
the year. This high-resolution data offers insight into daily consumer behaviour and seasonal
demand variations, which is essential for generating accurate inventory forecasts. Additionally,
the dataset includes daily opening inventory levels for each category, recorded at the
commencement of each business day. This inventory data reflects the stock available in the
warehouse and is crucial for analysing stock turnover rates, identifying potential stock shortages,
and planning replenishment cycles.
Delivery Data Compilation
Delivery logistics are tracked through data that encapsulates the journey of goods from suppliers
to the retailer's warehouse. It includes detailed logging of the average delivery time, measured
from when the supplier dispatches goods to their arrival at the warehouse. This metric is vital for
assessing the efficiency of the supply chain and identifying bottlenecks that may delay the
logistics process. Additionally, the dataset captures the average cost of delivering a unit of goods
from the supplier to the warehouse. This cost metric incorporates expenses such as fuel costs,
driver wages, and vehicle maintenance, providing a comprehensive understanding of the financial
aspects of delivery operations.
Environmental Influence Data
Weather conditions play a significant role in logistic operations, influencing both the delivery
speed and the products' integrity during transit. The dataset includes daily records of average
temperature and classifies the weather conditions into four categories: sunny, rainy, snowy, and
cloudy. These environmental data points are integrated into the AI models to predict and mitigate
potential disruptions caused by adverse weather conditions, ensuring robust delivery scheduling
and route planning that adapts to environmental changes.

Vol.13, No.2, May 2024
3
The primary objective of this research is to harness the capabilities of Artificial Intelligence (AI)
to enhance the logistical operations of retailers, focusing on inventory management and delivery
optimisation. Specific goals set for this study include:
Accurate Demand Forecasting: Develop predictive models that accurately forecast demand
across various product categories on a daily basis. This goal addresses the need to minimise
stockouts and overstock situations, which are critical for maintaining cost efficiency and
customer satisfaction.
Optimisation of Inventory Levels: Utilise insights gained from demand forecasting to establish
optimal inventory levels that ensure availability while minimising storage costs. This goal is
pivotal in reducing the capital tied up in inventory and the costs associated with storing excess
goods.
Enhanced Delivery Logistics: Improve the efficiency of delivery routes and schedules by
applying machine learning algorithms that consider multiple variables such as traffic patterns,
weather conditions, and geographical data. The aim here is to reduce delivery times and costs,
which in turn can enhance customer satisfaction and loyalty.
These objectives directly impact the practical aspects of retail operations as follows:
 Financial Performance: By reducing the incidence of overstocks and stockouts, retailers
can significantly decrease the costs associated with markdowns and lost sales, respectively.
Effective inventory management, driven by accurate demand forecasts, ensures that capital
is not unnecessarily tied up in inventory, improving the business's financial liquidity.
 Operational Efficiency: Streamlined inventory levels lead to more efficient use of
warehouse space and resources, including labour. Optimising delivery routes reduces fuel
consumption and wear and tear on delivery vehicles, further cutting operational costs.
 Customer Satisfaction: Ensuring product availability through accurate demand forecasts
and efficient inventory management prevents customer dissatisfaction due to stockouts.
Moreover, faster and more reliable delivery services directly contribute to a better
customer experience, fostering excellent customer retention and potentially increasing
market share.
3. METHODS
The methodology used in this research is a multifaceted approach that leverages both statistical
and machine-learning techniques to optimize retailers' logistical operations. The process begins
with applying time series analysis and extends through sophisticated machine learning models
and analytical techniques.
Time Series Analysis Using SARIMA
The research begins by applying the Seasonal Autoregressive Integrated Moving Average
(SARIMA) model to forecast initial demand. This model is essential for understanding the
dynamics of product demand over time by identifying patterns and seasonal variations in
historical sales data. This foundational analysis sets the stage for more detailed and predictive
analyses, providing initial insights into peak demand periods, critical for effective inventory
management.

Vol.13, No.2, May 2024
4
Enhanced Demand Forecasting with Random Forest
Following the initial time series forecasting, the study employs the Random Forest algorithm to
develop more comprehensive demand forecasts [13]. This model incorporates a broader array of
predictors, including promotional activities, economic conditions, and even weather changes,
beyond the fundamental trends identified by SARIMA. Random Forest, known for its robustness,
builds multiple decision trees to output a mean prediction that accounts for a wide range of
nonlinear dependencies and interactions between diverse variables.
Regression Analysis
In addition to time series and Random Forest models, regression analysis predicts quantitative
variables crucial to logistics operations, such as inventory levels and delivery times. This method
helps quantify the direct relationships between inputs (like sales forecasts and delivery schedules)
and outputs (such as inventory needs and delivery performance) [14], enabling precise
adjustments to operational tactics.
Cross-Validation
To ensure the models are accurate and generalizable, k-fold cross-validation is applied. It
involves dividing the data into several subsets, using each to test the model while training on the
remaining data [15]. This method is pivotal for verifying the model's effectiveness across
different scenarios and helps prevent overfitting.
Grid Search for Hyperparameter Tuning
Grid Search is implemented to fine-tune the hyperparameters of the Random Forest model, such
as the number of trees and their depth. This automated process tests various combinations of
parameters to determine the best configuration that optimizes a given performance metric
[16, 17]. This step is crucial to enhancing the model's predictive accuracy and efficiency.
Feature Importance Analysis
Finally, feature importance analysis within the Random Forest model identifies the variables that
significantly impact the forecasted outcomes. Understanding these key drivers is essential for
refining the predictive models and the logistical strategies they inform. This analysis helps
pinpoint critical factors that influence demand and delivery performance, guiding more focused
and effective inventory and delivery planning adjustments.
3.1. Interconnection of Research and Development Stages
The research and development stages are interconnected, starting from establishing a forecasting
base with time series analysis and the SARIMA model focusing on identifying and forecasting
main trends and seasonal fluctuations in product demand. These forecasts become the basis for
subsequent stages as they provide important information about future demand. The next stage is
the transition to multifactorial analysis, where data on trends and seasonality in demand obtained
in the first stage are integrated into regression analysis and random forest model analysis,
allowing for the consideration of additional factors such as weather conditions, holiday periods,
and logistic characteristics. It creates more accurate and comprehensive models for forecasting
key operational metrics. Further refinement and optimization of strategies follow, with forecasts
based on the SARIMA model providing an initial understanding of future goods needs.

Vol.13, No.2, May 2024
5
Regression analysis and the random forest model expand this analysis by accounting for a
broader set of influencing factors, more accurately determining optimal inventory levels and
planning more efficient delivery routes, thereby reducing costs and increasing customer
satisfaction. Regression analysis at this stage identifies general relationships between different
factors (weather conditions and distance to delivery points) and key operational metrics (time and
cost of delivery). After identifying critical relationships through regression analysis, data used at
this stage are integrated with additional data (historical sales data and inventory information),
allowing the random forest model to conduct a more comprehensive analysis. This integration
provides a broader context for analysis and optimization, considering many factors affecting
logistic operations.
From forecasting to strategic planning: Regression analysis provides forecasts based on identified
relationships, which are helpful for short-term planning and operational response to changes in
external conditions. The random forest model uses these forecasts and additional analytical data
to form longer-term optimization strategies, considering the variability and uncertainty of the
external environment. The final stage is prioritizing actions. Feature importance analysis
performed with the random forest model is directly linked to the regression analysis findings and
allows for identifying which factors should be given special attention when developing
operational and strategic plans [18, 19]. This information helps the retailer focus on the most
significant aspects for achieving maximum operational efficiency.
4. CALCULATIONS
In working with the SARIMA model, the following primary parameters were set: (2, 1, 2)x(1, 1,
1, 12), where: (2, 1, 2) indicates the autoregressive part (AR=2), order of integration (I=1), and
moving average order (MA=2) for the non-seasonal component; (1, 1, 1, 12) represents the
seasonal components of the model: seasonal autoregression (SAR=1), seasonal integration
(SI=1), seasonal moving average (SMA=1), and the season length of 12 months.
The Mean Absolute Error (MAE) was 2.5% of the average daily sales.
The Root Mean Squared Error (RMSE) was 3.2% of the average daily sales.
It means that the model accurately forecasts sales with minimal deviations from actual values.
The model also identified significant seasonal peaks in sales corresponding to holiday periods
and seasonal changes, such as increased sales in December and a decrease in January.
Additionally, an upward trend in annual sales was observed, indicating overall growth in demand
for the retailer's goods.
4.1. Forecasting Inventory Levels with Regression Analysis
The calculations yielded the following values: the Mean Absolute Error (MAE) was
approximately 6.67% of the average daily sales; the Root Mean Squared Error (RMSE) was
about 9.33% of the average sales volume.
4.2. Predicting Delivery Time
R-squared: 0.75, indicating that the used variables, including the distance to the delivery location,
weather conditions, and cargo volume explain 75% of the variability in delivery time. MAE: 0.5
hours, indicating that the average forecast error in delivery time is half an hour. RMSE: 0.65
hours, reflecting the standard deviation of forecast errors from actual data.

Vol.13, No.2, May 2024
6
During the research, analysis and modelling of key aspects of the retailer's logistic operations
were conducted using machine learning methods. Based on data representing sales dynamics,
inventory levels, and logistical costs over one calendar year, the following results were obtained:
4.3. Analysis of Variable Importance
Sales Historical Data. Importance: 42.7%. This feature is the most significant for forecasting
demand and, consequently, for optimizing inventory levels as it adapts to changes in consumer
preferences.
Weather Conditions. Importance: 18.3%. The impact of weather on delivery time is particularly
noticeable in extreme conditions, which can delay delivery or require route changes.
Distance to Delivery Location. Importance: 22.5%. The distance directly affects logistical costs
and delivery time, requiring efficient route planning.
Holiday Periods. Importance: 9.2%. Holiday periods often see an increase in demand, which
consequently changes logistic needs, including the need for additional inventory and delivery
plan adjustments.
Other Factors (including traffic congestion, type of vehicle, etc.). Importance: 7.3%. These
factors also contribute to optimizing delivery time and costs, albeit to a lesser extent than the
aforementioned features.
4.4. Delivery Time Metrics
The average delivery time was 24 hours, with the Mean Absolute Error (MAE) for delivery time
being 0.3 hours (18 minutes), approximately 1.25% of the average delivery time. The Root Mean
Squared Error (RMSE) for delivery time was 0.4 hours (24 minutes), approximately 1.67% of the
average delivery time. Regarding predicting delivery time, the model demonstrated an R-squared
of 0.75, confirming the ability to explain 75% of delivery time variations through the factors
used. Delivery time was predicted with an MAE of 0.5 hours and an RMSE of 0.65 hours,
indicating the model's efficiency in optimizing logistic processes.
5. RESULTS
This detailed analysis of the results derived from applying the Random Forest model in
forecasting and logistics optimisation focuses on the tangible benefits realised in the retailer's
operations. The comprehensive simulations and data analysis performed using the model have
facilitated precise demand forecasting and operational improvements in inventory management
and delivery efficiency.
The Random Forest model was applied extensively to predict future product demand across
various categories, employing a robust dataset incorporating historical sales, seasonal trends,
inventory levels, and external factors like market conditions. The predictions provided by the
model were instrumental in shaping inventory strategies that are responsive to actual market
demands, thereby minimising overstock and stockout scenarios.
AI-driven forecasts enabled a significant reduction in excess inventory, with the model achieving
a 20.5% reduction in surplus stock. This optimisation resulted in a corresponding decrease in
storage costs by 12.3%, highlighting the economic benefits of precise inventory forecasting. The

Vol.13, No.2, May 2024
7
model's effectiveness in aligning inventory with demand also led to a 14.8% decrease in stock
shortages, ensuring higher product availability and customer satisfaction.
Figure 1. Comparison of inventory management and delivery optimization metrics before and after
the application of analytical models
The model facilitated enhanced demand forecasting, which improved inventory turnover rates,
reflecting more efficient use of resources and capital. This is essential in industries where
products are subject to rapid obsolescence or seasonal variability. By maintaining optimal stock
levels, the retailer could avoid both capital lock-in and the need for emergency stock
replenishments, which are often costly and disruptive.
Further, the Random Forest model analysed logistics data, including store locations, historical
delivery patterns, traffic data, and weather conditions. This analysis was crucial in developing
optimised delivery routes and schedules contributing to operational efficiencies.
By implementing AI-recommended routes, the average delivery time was reduced by 12%
compared to traditional delivery methods. This improvement not only enhances customer service
by ensuring faster delivery but also increases the efficiency of the logistics network, allowing for
more deliveries to be completed within the same time frame.
The optimisation of delivery routes led to a noticeable reduction in delivery costs, evidenced by a
7.5% decrease. These savings arise from reduced fuel consumption, lower vehicle maintenance
costs, and improved utilisation of delivery personnel. The strategic routing enabled by the model
minimises unnecessary travel, thereby reducing wear and tear on vehicles and diminishing the
environmental impact of delivery operations.
5.1. Analysis of Variable Importance
Sales Historical Data: The analysis underscored their key role in forecasting future demand with
an accuracy of up to 92.3%, reinforcing the position of collecting and analyzing sales data as a
basis for inventory planning.

Vol.13, No.2, May 2024
8
Weather Conditions and Holiday Periods: Significantly impacted delivery time and demand,
contributing to a 15.2% change in delivery time and affecting demand by 13.4%, confirming the
necessity of considering these factors in logistic operation planning.
6. CONCLUSION AND DISCUSSION
This article presents data and results on various analysis models for optimizing a retailer's logistic
operations. All data are coherent and logically interconnected. The SARIMA model and its
accuracy in sales forecasting, as indicated by low MAE and RMSE values, demonstrate its ability
to predict daily sales and identify seasonal sales peaks. It matches expectations for a model
designed to analyze time series with seasonal fluctuations. The regression analysis results for
forecasting inventory levels and delivery times also align with the realistic expectations from
such analytical models. The MAE and RMSE metrics for inventory level and delivery time
forecasts underscore that the model provides relatively accurate predictions, essential for
effective inventory management and logistics. The variable importance analysis offers valuable
insights into which factors most significantly impact inventory levels and delivery times.
Historical sales data, weather conditions, distance to the delivery location, and holiday periods
are important regarding their impact on logistic processes. Integrating data and models for
comprehensive operation optimization shows how different methods can complement each other,
providing a holistic approach to inventory and logistics management. Thus, the outcomes of
various models and analysis methods present a logical and consistent picture, where each element
contributes to the overall goal of optimizing the retailer's logistic operations.
The practical implementation of the described methods involves several key steps and tools.
Initially, an integration platform capable of collecting data from various sources (CRM, ERP,
external weather APIs, logistics services) and processing it in one place is required. It could be
specialized business analytics software or a custom system built on cloud services (e.g., AWS,
Google Cloud, Azure).
Automating data collection necessitates using scripts or ETL processes (Extract, Transform,
Load) to automatically extract data, transform it into the required format, and load it into an
analytical database. It can be achieved with tools like Apache NiFi, Talend, or built-in cloud
platform instruments.
Analytical models based on SARIMA and random forest are constructed and trained using
programming languages that support statistical and machine learning – such as Python or R.
Libraries like pandas for data processing, scikit-learn for machine learning, and statsmodels for
statistical models are used.
For aggregating results and making decisions, business logic based on rules or algorithms
integrates findings from different models to form final recommendations. This could be
implemented through a comprehensive software module written in Python, which analyzes model
outcomes, compares them, and presents optimal suggestions.
Data visualization tools like Tableau, Power BI, or cloud services' built-in solutions (Google Data
Studio, AWS QuickSight) can be used to create interactive dashboards displaying analytics,
forecasts, and recommendations in an understandable format. To automate reporting and
notifications, developing a notification system that can automatically inform managers about
significant changes or action requirements via email, SMS, or integration with corporate
messengers (Slack, Microsoft Teams) is advisable. Implementing feedback mechanisms to collect

Vol.13, No.2, May 2024
9
data on the effectiveness of decisions and their impact on business processes is crucial for further
model and business logic improvement.
Creating such a comprehensive system requires significant investment and time to become viable
and applicable in practice. Besides building the system's structure, having a necessary and
sufficient data set is crucial for its effectiveness. However, the study has shown that applying a
comprehensive analytical approach to managing a retailer's logistic operations offers substantial
advantages, enhancing the efficiency of current operations and providing a basis for making
informed strategic decisions. Integrating diverse data and analytical models is key to successfully
solving complex logistic challenges and ensuring sustainable growth and competitiveness in a
dynamic business environment.
REFERENCES
[1] Y. Chen and X. Zheng, (2024) "Optimization strategy of cross-border e-commerce supply chain
network based on machine learning," Applied Mathematics and Nonlinear Sciences, Vol. 9, No. 1.
[2] I. A. Shah, R. K. Murugesan, and S. Rajper, (2024) "Supply Chain Management Security Issues and
Challenges in the Context of AI Applications," in Navigating Cyber Threats and Cybersecurity in
the Logistics Industry, pp. 59-89.
[3] T. Archana and R. K. Stephen, (2024) "The Future of Artificial Intelligence in Manufacturing
Industries," in Industry Applications of Thrust Manufacturing: Convergence with Real-Time Data
and AI, IGI Global, pp. 98-117.
[4] S.-W. Lin and W.-M. Lu, (2024) "Using inverse DEA and machine learning algorithms to evaluate
and predict suppliers’ performance in the apple supply chain," International Journal of Production
Economics, 109203.
[5] A. Khan and A. Ahmed, (2024) "Optimizing Retail Operations, Inventory Management and Sales
Forecasting with Big Data and AI in China," Emerging Trends in Machine Intelligence and Big
Data Engineering.
[6] S. Hasnaoui, (2023) "New Trends in Smart Supply Chain Management: An LSTM-Based Solution,"
Repository ESI-SBA.
[7] R. Couronné, P. Probst, and A.-L. Boulesteix, (2018) "Random forest versus logistic regression: a
large-scale benchmark experiment," BMC Bioinformatics, Vol. 19, pp. 1-14.
[8] S. Choudhary et al., (2021) "Innovative supply chain solution for Indian automobile manufacturing
sector: a case study," EPRA International Journal of Research and Development (IJRD), Vol. 6, No.
6.
[9] I. A. Shah, R. K. Murugesan, and S. Rajper, (2024) "Supply Chain Management Security Issues and
Challenges in the Context of AI Applications," in Navigating Cyber Threats and Cybersecurity in
the Logistics Industry, pp. 59-89.
[10] Nebojša Vasić et al., (2021) "Satisfaction is a function of users of logistics services in e-commerce,"
Technology Analysis & Strategic Management, vol. 33, No. 7, pp. 813-828.
[11] A. E. Laksana and B. J. Santoso, (2023) "Predicting Passenger Occupancy of Commercial Buses
Using Regression Approach of Machine Learning," in Proceedings of the 2023 International
Conference on Advanced Mechatronics, Intelligent Manufacture, and Industrial Automation
(ICAMIMIA), IEEE, pp. 988-993.
[12] K. Sreenivasulu, B. Sowjanya, V. R. Motupalli, S. H. Yadav, K. K. Baseer, and M. J. Pasha, (2023)
"Prediction of Flight Delay Through Intelligent Algorithms and Big Data Technology," in
Proceedings of the 2023 2nd International Conference on Applied Artificial Intelligence and
Computing (ICAAIC), IEEE, pp. 1074-1080.
[13] H. Nassiri, S. I. Mohammadpour, and M. Dahaghin, (2023) "Forecasting time trends of fatal motor
vehicle crashes in Iran using an ensemble learning algorithm," Traffic Injury Prevention, Vol. 24,
No. 1, pp. 44-49.
[14] P. Hatami Bahman Beiglou, L. Luo, P. N. Tan, and L. Pei, (2021) "Automated analysis of the US
Drought Monitor maps with machine learning and multiple drought indicators," Frontiers in Big
Data, Vol. 4, 750536.

Vol.13, No.2, May 2024
10
[15] Y. N. Fuadah, M. A. Pramudito, and K. M. Lim, (2022) "An optimal approach for heart sound
classification using grid search in hyperparameter optimization of machine learning,"
Bioengineering, Vol. 10, No. 1, pp. 45.
[16] S. Chaudhury, N. Shelke, Z. M. Rashid, and K. Sau, (2022) "Effect of Grid Search and Hyper
Parameter Tuned Pipeline with Various Classifiers and PCA for Breast Cancer Detection," Current
Signal Transduction Therapy, Vol. 17, No. 3, pp. 45-56.
[17] B. S. Vardhan, M. Khedkar, and P. Thakre, (2022) "A Comparative Analysis of Hold Out, Cross
and Re-Substitution Validation in Hyper-Parameter Tuned Stochastic Short Term Load
Forecasting," in Proceedings of the 2022 22nd National Power Systems Conference (NPSC), IEEE,
pp. 448-453.
[18] R. Fekadu, A. Getachew, Y. Tadele, N. Ali, and I. Goytom, (2022) "Machine Learning Models
Evaluation and Feature Importance Analysis on NPL Dataset," arXiv preprint arXiv:2209.09638.
[19] H. Chung, H. Ko, W. S. Kang, K. W. Kim, H. Lee, C. Park, ... & J. Lee, (2021) "Prediction and
feature importance analysis for severity of COVID-19 in South Korea using artificial intelligence:
model development and validation," Journal of Medical Internet Research, Vol. 23, No. 4, e27060.

OPTIMISATION OF LOGISTIC OPERATIONS USING AI

More Related Content

Similar to OPTIMISATION OF LOGISTIC OPERATIONS USING AI (20)

Recently uploaded (20)

OPTIMISATION OF LOGISTIC OPERATIONS USING AI