Multi-echelon distribution network inventory optimization for cross-border omni-channel e-commerce

In a company, inventory management is crucial due to the significant impact on various aspects of the business. Similarly, the Indonesian water supply company (PDAM) requires effective inventory management to ensure the supply of liquid aluminum sulfate chemicals. The probabilistic statistical inventory control (SIC) model is commonly used for inventory management. However, previous research on chemical inventory models in PDAMs often relied on simple linear regression to forecast demand data, which fails to capture the inherent volatility in demand. Therefore, this research aimed to predict demand data using the seasonal autoregressive integrated moving average (SARIMA) method and determine the optimal policy for supplying liquid aluminum sulfate chemicals. The results showed that the best demand forecasting model was SARIMA (2,1,2) (1,1,0)12 with a mean absolute percentage error (MAPE) value of 8.19%. The finding of the optimal inventory policy showed a safety stock value of 11,922.35 kg, a reorder point value of 49,511.20 kg, and an order quantity of 21,526.59 kg, leading to a total cost of IDR 11,132,034,145.45. The sensitivity test also showed that variations in lead time, price, μ, and σ parameters directly influence changes in total cost, reorder point, and safety stock. These calculations were conducted using Minitab and Python software.

With the continuous expansion of cross-border e-commerce enterprises' cross-border business activities and the rising operating costs of overseas warehousing, the cross-border e-commerce enterprise resource management system (ERP) is responsible for the management, coordination and optimization of global supply chain warehousing. On the cross-border e-commerce ERP platform, the product sales forecast and inventory optimization strategy realized by machine learning algorithm can effectively summarize the key factors, use the sales record big data, so that the forecast value, the expectations fit the actual value of the basic trend. The results show that the machine learning algorithm has good prediction effect and improved the best inventory equilibrium efficiency.

Currently all companies are looking for opportunities in the openness of international markets or global markets. Companies that can enter this global market are companies that have comparative and competitive advantages. It is necessary to make an important decision concerning the level or quantity of production for the manufacturing company, the determination of the amount or stock of merchandise for the trading company as well as the level of service production for the service company. One important aspect that needs proper management is the supply problem. One of the local companies that became the focus or object of research in this writing is UD. Mirama located in Gorontalo Province, especially the city of Gorontalo which sells AC electronic goods and TV. just like any other trading company the company is also having trouble determining the amount of merchandise inventory. Because often companies have difficulty in the amount of safety stock that must be prepared at the time of booking so that companies can anticipate the number of fluctuating demand. With this problem it is necessary to apply a method used as corporate guidance to overcome the problem of inventory. The method that will try to use is standard deviation method in determining the amount of safety stock. This method is applied to see how the most optimal security inventory, which is the most economical, in the sense that not too much which means waste or additional costs that are not necessary or not too little which means there is still danger of running out of inventory. By using model of qualitative and quantitative analysis, the result of research analysis from the research object is the application of standard deviation method can be determined how much the safety stock will be held by the company and can cover the occurrence of stock out when demand increase or fluctuation in demand.

Multi-echelon supply chain networks face increasing complexity in demand forecasting and inventory optimization due to volatile market conditions and dynamic customer preferences. Traditional forecasting methods often struggle to capture long-range dependencies and complex temporal patterns inherent in supply chain data. Transformer-based architectures, originally developed for natural language processing (NLP), have emerged as powerful tools for time series forecasting in supply chain management (SCM). These models leverage self-attention mechanisms to process sequential data and capture intricate relationships across multiple time steps. This review examines the application of transformer models in demand forecasting and inventory optimization within multi-echelon supply chain networks. The paper analyzes how transformer architectures address challenges such as bullwhip effect amplification, demand volatility, and coordination across supply chain tiers. Key findings indicate that transformer-based approaches outperform conventional methods including autoregressive integrated moving average (ARIMA), long short-term memory (LSTM) networks, and traditional machine learning (ML) algorithms in forecast accuracy and computational efficiency. The review synthesizes recent developments in transformer architectures specifically adapted for supply chain contexts, including modifications for handling sparse data, incorporating external factors, and enabling real-time decision support. Furthermore, the paper explores integration strategies between demand forecasting and inventory optimization, examining how transformer predictions inform safety stock calculations, reorder point determination, and dynamic replenishment policies. Emerging trends such as attention mechanism interpretability, federated learning (FL) for collaborative forecasting, and hybrid models combining transformers with reinforcement learning (RL) are discussed. The review identifies critical gaps in current research, including limited validation in real-world multi-echelon settings, computational scalability challenges, and the need for robust frameworks addressing demand uncertainty propagation across supply chain levels.

The Capacitated Vehicle Routing Problem with Time-Dependent Demands (CVRPTD) is a significant optimization challenge in the logistics and transportation domain, characterized by dynamic customer demands, strict time windows, and heterogeneous vehicle fleets. This study focuses on urban parcel delivery operations as the primary object of research. The problem addressed involves the inefficiency of conventional vehicle routing strategies in adapting to time-varying customer demands and operational constraints, which often lead to increased costs and service delays. This study aims to minimize total operational costs while ensuring compliance with capacity constraints, service continuity, and demand fluctuations. A comprehensive mathematical model is developed based on a fully connected, directed acyclic graph G=(V, A), incorporating decision variables that represent vehicle routing sequences, timing, and vehicle type assignments. This study addresses the Capacitated Vehicle Routing Problem with Time-Dependent Demands (CVRPTD) in urban parcel delivery, where traditional routing methods struggle with dynamic demands and operational constraints. A mathematical model using a directed acyclic graph is developed, optimized via a gradient-based method with Hessian approximation, LU decomposition, and quasi-Newton techniques. Experiments on datasets with up to 200 customers and 20 vehicles with reductions ranging from 1.79% to 12.75%. The most significant improvement was observed in Sidorame Timur, where the optimization distance decreased by 12.75%, indicating high accuracy in route optimization. For the SCP, the proposed algorithm achieved a 6.46% improvement in solution quality over traditional greedy algorithms

Increasing environmental considerations force the logistics sector to develop more environmentally friendly approaches. Thus, the concept of "green logistics" is becoming more crucial. The use of electric vehicles, which become prominent with their low energy consumption, less cost and environmentally friendly features, is a part of green logistics. Moreover, taking into consideration of multi-echelon distribution networks can also be effective in reducing the negative outcomes of urban logistics. In this context, the Two-Echelon Electric Vehicle Routing Problem with Battery Swapping Station (2E-E-VRP-BSS) is considered in this study. A Mixed Integer Programming (MIP) model is presented for the problem in which the total cost consisting of fixed cost, transportation cost, battery swapping cost and handling cost is minimized. To test the validity of the model, three different data sets, each of different sizes, simulating a realistic distribution network were created. In addition, sensitivity analyses were carried out in order to examine the impact of changes in customer demand and battery capacities of electric vehicles on optimal solutions. The analyses indicate the applicability of the model. The findings suggest that fluctuations in customer demand directly impact the total cost, with higher demand leading to increased costs and lower demand resulting in cost reductions compared to the base case scenario. Furthermore, there is a negative correlation between the battery capacity of electric vehicles and the total cost, meaning that as battery capacity increases, the overall cost tends to decrease.

Supply chain waste manifests through multiple mechanisms including excess inventory accumulation, stockouts leading to lost sales and expedited fulfillment costs, product obsolescence in dynamic markets, and systematic forecast errors that propagate throughout distribution networks. These inefficiencies impose substantial operational and financial burdens while undermining sustainability objectives through resource misallocation and increased carbon emissions. This article examines how artificial intelligence-enabled demand forecasting and inventory optimization can systematically reduce waste across multi-tier supply chains. Advanced AI techniques including probabilistic forecasting methods, deep learning architectures for pattern recognition in complex demand signals, and reinforcement learning for dynamic policy adaptation enable more accurate demand predictions and responsive inventory positioning. Multi-echelon inventory planning frameworks optimize stock allocation across distribution layers, while safety stock optimization balances service-level requirements against holding costs under uncertainty. Dynamic replenishment strategies leverage real-time data streams to adjust ordering decisions continuously. Empirical evidence demonstrates that these approaches yield substantial cost reductions through decreased holding and shortage costs, improved service-level attainment through better availability, and enhanced sustainability outcomes via reduced waste generation and resource consumption. Future directions include the development of explainable AI systems that provide transparent decision rationale, real-time decision intelligence platforms that integrate sensing and actuation capabilities, and explicit consideration of ethical dimensions in automated planning systems including fairness, accountability, and human oversight requirements.

Subject of study. Purpose of the research: The aim of the study is to develop a two-stage supply chain management model that allows for effective responses to market changes and enhances the adaptability of enterprises. Special focus is placed on reducing risks and optimizing resources at each stage of the supply chain. Research methods: The study uses theoretical models of supply chain management at both strategic and operational levels, as well as numerical methods for process optimization. To assess the effectiveness of the proposed model, mathematical modeling and computer simulation methods are employed, which account for dynamic changes in supply conditions and demand fluctuations. Results of the work: The influence of the two-stage model is assessed in the context of financial indicators, specifically on the capital turnover, return on investment and liquidity. It was discovered that effective usage of two-stage structurization contributes to logistics cost shrinkage, inventory optimization and increase of financial stability of the enterprise. It was stated that effective reacting on market instability is possible under the condition of adaptive management and flexible contract mechanisms usage, which can also contribute to operational costs optimization and profitability levels increase. Potential research can focus on forecasting analytical models integration for the purpose of improving the accuracy of made decisions in complex market conditions.

This study focuses on the optimization of spare part inventory for mask production machines at PT. XYZ with the aim of enhancing productivity and reducing operational costs. The research identifies key failure points in the mask machines, including issues with the conveyor system, pulley pulls, and ultrasonic sealers, which were the main contributors to production downtime. This study adopts a quantitative research approach, focusing on the systematic collection and analysis of numerical data to draw conclusions and provide objective insights into the issues being studied. By applying inventory management concepts such as safety stock, service levels, and periodic reviews, the study determines the optimal spare part inventory levels required to ensure component availability and minimize the risk of stockouts. The implementation of the PDCA (Plan-Do-Check-Act) method further improves machine maintenance processes, leading to an increase in machine effectiveness levels. Additionally, effective inventory management reduces waste due to over-purchasing and mismanagement of spare parts. The findings highlight that by optimizing spare part inventory and improving machine maintenance, PT. XYZ was able to reduce downtime, minimize operational costs, and enhance overall productivity, contributing to greater efficiency and profitability.

This study aims to strengthen the participation of Small and Medium Enterprises (SMEs) in government-led construction supply chains by examining three key dimensions: material inventory optimization, identification of participation barriers, and actionable strategies for inclusion. The research was conducted in Surkhet Valley, Nepal, a region actively engaged in public infrastructure development. For the first objective, a three-year dataset of publicly procured construction materials was analyzed using ABC analysis based on usage value and frequency. Results revealed that five materials, including reinforcement and cement, accounted for 75% of total expenditure, highlighting the need for focused inventory control on high-value items. To address the second objective, perceptions from 58 SMEs were analyzed using Principal Component Analysis (PCA), identifying three major factors that explain 68.47% of the variance in SME challenges: (1) structural and operational barriers, such as bureaucratic delays; (2) financial and logistical constraints, like payment delays and demand volatility; and (3) capacity-related issues, including limited training and weak quality systems. For the third objective, the Relative Importance Index (RII) and Key Informant Interviews (KIIs) were used to prioritize remedial measures. Top-ranked strategies included cost minimization, simplified procurement paperwork, and improved logistics. KIIs further emphasized the importance of transparent payment systems and SME-friendly tender policies. The study concludes that a focused approach grounded in material classification, systemic barrier reduction, and SME-centric policy design can strengthen SME participation in public construction supply chains. Recommendations include ABC-based procurement planning, simplifying documentation requirements, SME training, and establishing preferential procurement policies.

Stock management at PT. Gasindo Raya Utama is a system designed to simplify stock management and facilitate accurate decision-making processes. The purpose of this system is to design an information system that supports the management of 3 kg LPG stocks at PT. Gasindo Raya Utama by implementing the Safety Stock and Reorder Point methods. This system is intended to help the company prevent stock shortages and determine the timing and quantity of reorder effectively. The Safety Stock method is applied to calculate the minimum reserve stock required to anticipate fluctuations in demand or supply delays. Meanwhile, the Reorder Point method is used to determine the reorder point based on daily consumption levels and lead times. The implementation of this system is expected to improve the efficiency of stock management, minimize the risk of stockouts, and ensure that LPG distribution runs smoothly in accordance with the needs of the community.

The efficient distribution of pharmaceutical products is crucial for ensuring timely access to essential medications, particularly in resource-constrained environments, while minimizing costs and waste. Traditional supply chain approaches often struggle with demand volatility, transportation constraints, and suboptimal inventory management. This research explores the application of reinforcement learning (RL) to optimize drug distribution in pharmaceutical logistics. We propose an RL-based framework that leverages real-time data, including GPS tracking, warehouse inventory levels, and demand fluctuations, to enhance adaptive decision-making. Using Deep Q-Networks (DQN), the model dynamically optimizes routing, inventory restocking, and demand forecasting to minimize delivery delays and reduce holding costs. Experimental results demonstrate significant improvements in delivery efficiency, reduction in stock shortages, and enhanced overall supply chain performance compared to conventional logistics strategies. This study highlights the potential of RL-driven approaches in transforming pharmaceutical distribution, ensuring equitable and timely medication availability.

Multi-tier goods inventory strategy and optimization in large enterprises are hard nuts to crack. Thus goods distribution network often either carry excess total inventory, or lower customer service level due to improper goods safety stock in the network. Two typical approaches for managing multi-tier inventory and their pitfalls are examined in this study. Based on the review, real multi-tier inventory strategy and optimization tools are presented with a simulation result to reflect the impact of multi-tier strategy on the network inventory optimization. The result shows that uncertainties in supplier delivery lead-time and/or demand variations will be the two most important factors with non-linear impact.

Control model for spare parts inventory is established based on the optimal replenishment cycle. The replenishment cycle impact on the spare parts inventory control is analyzed as well as the demand forecasting. First, the optimal replenishment cycle is given by using the method of the lowest cost of inventory. Then according to the demand characteristic of the spare parts, the safety stock can be calculated. Finally on the basis of the demand forecasting, the calculation method of spare parts replenishment quantity is given. A numerical example is presented to verify the validity and practicability of the model.

The research observed the Supply Chain (SC) of Ammonium nitrate (AN), a critical and high-value explosive material in coal mining operation, which faces disruptions due to fluctuations of global and domestic coal demand, impacting to AN demand volatility. The lengthy, complex, dependency on sole suppliers, and non-integrated business process, coupled with specific mining conditions, results in long procurement lead times, high inventory costs, and coal production halt. This research used Kepner-Tregoe Analysis (KTA), Business Process Engineering (BPE), Vendor Managed Inventory (VMI), and Economic Order Quantity (EOQ) methods to appraises the complex SC situation, assess the business processes, develop an effective and efficient inventory management model and quantify the cost analysis. The VMI design will streamline SC processes by segregating the responsibilities: vendor manages daily procurement and warehouse operations, while the company provides oversight. It also aims to minimize unpredictable delivery times, elimination of shortages through optimized vendor stock, price and fall pricing stability via long-term agreements, enhanced cash flow through a monthly invoice, usage-based reconciliation, and a significant reduction in ordering and inventory holding costs. Eventually, this research contributes the benefits to Company, vendor, and as references for SC experts to develop a strategic SC Management through VMI implementation.