Applying machine learning to the dynamic selection of replenishment policies in fast-changing supply chain environments

Firms currently operate in highly competitive scenarios, where the environmental conditions evolve over time. Many factors intervene simultaneously and their hard-to-interpret interactions throughout the supply chain greatly complicate decision-making. The complexity clearly manifests itself in the field of inventory management, in which determining the optimal replenishment rule often becomes an intractable problem. This paper applies machine learning to help managers understand these complex scenarios and better manage the inventory flow. Building on a dynamic framework, we employ an inductive learning algorithm for setting the most appropriate replenishment policy over time by reacting to the environmental changes. This approach proves to be effective in a three-echelon supply chain where the scenario is defined by seven variables (cost structure, demand variability, three lead times, and two partners’ inventory policy). Considering four alternatives, the algorithm determines the best replenishment rule around 88% of the time. This leads to a noticeable reduction of operating costs against static alternatives. Interestingly, we observe that the nodes are much more sensitive to inventory decisions in the lower echelons than in the upper echelons of the supply chain.     

Revisiting local TOC measures in an internal supply chain: a note

While the global theory of constraints (TOC)-based measures of throughput, inventory and operating expense are well known, the literature on local measures of throughput dollar-days and inventory dollar-days (TDDs/IDDs) is sparse and inconclusive. This paper shows how adherence to these measures naturally puts a non-TOC company on a continuous improvement path by assisting in identification of the constraint and implementation of a drum–buffer–rope (DBR)-like system. Thus, by simulating a small company, this note highlights the relationship between TDD/IDD measures and a DBR system, which further promotes TOC principles. Ultimately, we provide future research avenues into the role these measures can play in the development of a holistic incentive system and the implementation of a DBR-like system across the supply chain network.     

Evaluating the effects of capacity constraints and demand patterns on supply chain replenishment strategies

This research considers inventory replenishment in a stochastic, multi-echelon supply chain involving both production and distribution functions. Simulation is used to compare distribution/material requirements planning (DRP/MRP), re-order point (ROP) and Kanban (KBN) replenishment strategies. Additional experimental factors include the demand pattern and the existence of manufacturing capacity constraints. Trade-off curves between inventory and delivery performance are generated. Statistical techniques, including analysis of variance (ANOVA), are then used to compare the areas under the trade-off curves and determine the relative dominance among the replenishment strategies. The methodology is used to identify both main and interaction effects. With seasonal demand, DRP/MRP performance is found to be best, followed by ROP and KBN, respectively. Without seasonal demand, the relative performance ranking depends on the presence of capacity constraints. Without capacity constraints, ROP performs best, followed by DRP/MRP and KBN. With capacity constraints, the ranking is reversed. This difference in behaviour can be explained using queuing analysis.     

Identifying inventory problems in the aerospace industry using the theory of constraints

In the aerospace industry, where each component is time-critical, many operation management tools are applied to improve material management efficiency. By using the theory of constraints (TOCs), this study diagnoses the undesirable effects (UDEs) of material management systems and identifies root problems or conflicting material management actions. In the second step, via UDEs, this study constructs a current reality tree (CRT) to identify a company's material management objectives and requirements, and the actions it takes to meet these objectives. In addition, the UDEs and CRT are applied to redesign the future reality tree (FRT) to develop strategies and eliminate the problems in the case company. By eliminating the conflicting inventory management activities of different departments, inventory management performance for the entire system is improved.     

The bullwhip effect in hybrid supply chain

As the service is becoming the strategic preference in the manufacturing industry, more and more manufacturing companies provide customers with total solutions by integrating services into their core product offerings, which is usually implemented in so-called hybrid supply chain (HSC). The HSC is formed as a chain by a group of manufacturers and service providers who work together to offer total solutions to customers. In HSC, customers’ responses to variable service levels and the interaction between inventory strategy and service capacity strategy have a great impact on supply chain instability, which represents an important concern in supply chain research. To tackle this problem, we establish a system dynamics model, by taking the HSC for an elevator company in China as an example. Firstly, we analyzed the oscillation characteristics of service flow and product flow. Then, we proposed the performance metrics of bullwhip effect in HSC. Finally, based on the model, we find that the bullwhip effect of HSC could be smoothened by incorporating service capacity adjustment strategy into the inventory replenishment policy.     

Divergence control of a one-level supply chain replenishment rule

We present a general control and improvement strategy for a one-level supply chain based on maintaining the divergence of the system close to zero at each time step. The online implementations as well as the results obtained are shown for a logistic chain model with an Order-Up-To policy using several demand patterns. The divergence can be obtained using the state-space volume calculated with all the state variables of the model. However, it is also possible to calculate the divergence by applying state-space reconstruction techniques using only one state variable. The results obtained with both approaches show that this strategy allows the reduction in the total cost.     

Mapping the future of supply chain management: a Delphi study

Supply chain management, a field that developed from business practice and research, is undergoing a major transformation. It is changing from tactical in nature (where the major focus is on cost and delivery) to a field that is strategic in nature. However, the future issues and challenges facing managers and executives are just now becoming understood. This paper reports these issues by drawing on the findings generated by a three-phase study consisting of a literature review, a two-round Delphi study, and a workshop. Unique in this Delphi study is that it brings together leading practitioners in supply chain management with leading supply chain management researchers. The findings show that while the focus of the current tactical supply chain view is relatively limited to issues of delivery, risk, and leadership, the supply chain view of the future (i.e., five years from now) is more complex and demanding. The findings also show that there is generally no difference between researchers and practitioners in terms of how they view the issues. Finally, the study uncovers major obstacles that must be resolved before the strategic potential of future-state supply chains can be realised.     

Inventory management in a lateral collaborative manufacturing supply chain: a simulation study

Due to global competition, firms are seeking more effective supply chain (SC) collaboration in order to provide quality products with less cost, at the right time and in the right quantity. The present study examines manufacturing SC collaboration on the basis of holding cost, backorder cost and ordering cost. The types of collaboration examined are vertical, horizontal and lateral collaboration. This research emphasises lateral collaboration by determining the impact of inventory policies ((s, S) and (s, Q) inventory policies) on SC performance. For better understanding, a conceptual model is provided that is supported by a numerical example. As the study of SCs is complex in nature, a simulation approach has been employed to show the impact of lateral collaboration on performance measures such as the total cost, which is the sum of several cost components: inventory holding cost, backorder cost and ordering cost. The research is based on two manufacturing SCs where the manufacturer is taken as the collaborative node. To allow more clarity, a separate study on each cost component has been conducted. The laterally collaborative SC was simulated on ARENA 9.0, a simulation package. The results show that the efficacy of lateral collaboration outperforms horizontal collaboration due to having the individual SC members at more liberty to make decisions.     

Consumer search,transshipment, and bargaining power in a supply chain

We study a decentralised supply chain with one manufacturer and two symmetric retailers who could transship their excess inventory between each other. Without transshipment, the retailer’s unsatisfied customers search and buy the product at the other retailer with some probability. With transshipment, the retailer can transship the other retailer’s excess inventory to fulfil its own unmet demand. However, a transshipment price is often charged by one retailer for transshipping goods to another. We show the retailer’s profit is unimodal in the transshipment price, and the manufacturer’s profit increases in the customer search probability. Although the retailer’s preference over transshipment depends on the magnitude of search probability and transshipment price, the retailer may always prefer the transshipment no matter their magnitude. We then study the effect of bargaining power over the transshipment decision (whether to transship between retailers and at what transshipment price). Our analysis suggests that the manufacturer always prefers to control the transshipment price rather than to control whether to transship, while the retailer can prefer both. We also find that increasing bargaining power always benefits the manufacturer but could hurt the retailer. We finally check the robustness of our results by investigating the case with asymmetric retailers     

Roles of inventory and reserve capacity in mitigating supply chain disruption risk

This research focuses on managing disruption risk in supply chains using inventory and reserve capacity under stochastic demand. While inventory can be considered as a speculative risk mitigation lever, reserve capacity can be used in a reactive fashion when a disruption occurs. We determine optimal inventory levels and reserve capacity production rates for a firm that is exposed to supply chain disruption risk. We fully characterise four main risk mitigation strategies: inventory strategy, reserve capacity strategy, mixed strategy and passive acceptance. We illustrate how the optimal risk mitigation strategy depends on product characteristics (functional versus innovative) and supply chain characteristics (agile versus efficient). This work is inspired from a risk management problem of a leading pharmaceutical company.     

制约理论之思维流程在某制造企业的应用

按照思维流程的三条逻辑脉络:改变什么,改成什么样,怎样去改.以某制造企业为例说明了TP的五大逻辑关系图:当前现实树、冲突化解图、未来现实树、必备树和转移树的运用过程,以帮助企业发现核心问题,找到解决问题的方案并采取行动最终解决问题.     

Lateral inventory transshipment problem in online-to-offline supply chain

Online-to-offline (OTO) is a new commercial model with enormous market potential. Online customer orders are forwarded to the offline brick-and-mortar store to fulfil, which is a combination of dual-channel supply chain. OTO overcomes many disadvantages of the traditional dual-channel supply chain, but still faces uncertain market demand. To reduce the inventory risk caused by demand uncertainty, lateral inventory transshipment is employed in this paper to pool inventory risk in OTO supply chain. We model centralised OTO and decentralised OTO with/without transshipment, and then analyse different scenarios. Our results demonstrate that there exists a unique Nash equilibrium of inventory order levels in dual channels and an optimal transshipment price to maximise the profit of the entire supply chain. Finally, we provide a numerical example of uniform demand distribution. Our analyses offer many managerial insights and show that transshipment always benefits the OTO supply chain.     

How SMEs can benefit from supply chain partnerships

In recent literature on supply chain partnerships in small and medium-sized enterprises (SMEs), there is controversy regarding the benefits of these partnerships. To resolve this controversy, specific information is needed on the implementation of these partnerships by SMEs; an area, that, thus far, has received little academic attention. In this paper, we examine different business functions (production, marketing and sales, purchasing and logistics, research and development (R&D) and finance) within a supply chain partnership. We collected data for each individual function from 279 high-tech SMEs and examined the relationship between the specific types of partnerships and the overall performance of the SMEs. The results indicate that it is only in the area of R&D that partnerships have a significant positive effect on overall firm performance. The results imply that SMEs primarily can benefit from particular types of supply chain partnerships, i.e. R&D partnerships. The results contribute to the debate in the literature by explaining why many SMEs were found not to benefit from these partnerships. We also provide implications for firms and how SMEs can better utilise supply chain management (SCM).     

Throughput/inventory dollar-days: TOC-based measures for supply chain collaboration

As the semiconductor industry moves away from vertical integration, performance measures play an increasingly important role to ensure effective collaboration. This paper demonstrates that the theory of constraints (TOC)-based measures, Throughput and Inventory Dollar-Days (T/IDD), induce autonomous supply chain (SC) links to function as a synergistic whole and thereby, improve the performance of the whole SC network significantly. We model an SC network of a well-known TOC case study using discrete event simulation and discuss managerial implications of these measures via a set of scenarios. The scenarios explain how these measures – without sharing sensitive financial data – allow members of an SC network to monitor both the effectiveness (TDD) and efficiency (IDD) of SC members and lead them to create win-win solutions following well-known TOC-based planning and control concepts. We conclude this paper by discussing some limitations of the proposed research and provide directions for future theoretical research.     

Bottleneck identification in supply chain networks

In supply chain risk management, it is essential to identify firms that induce high losses due to supply chain disruptions in a focal firm or the supply chain network as a whole (bottlenecks). In this article, we describe supply chain networks as complex systems of firms and their suppliers. We revisit some established network measures and compare their predictions with a new methodology for detecting bottlenecks. In this bottom-up approach, production disruptions on the firm level are modelled with stochastic point processes, and a mechanism for the propagation of losses through the network is defined. The individual firms' emerging loss contributions to the total losses of the focal firm provide, then, an alternative risk-adjusted measure. Our methodology and findings enable more informed and transparent decisions to be made for optimal supply chain network design.     

Stockout propagation and amplification in supply chain inventory systems

We investigate the existence and magnitude of stockout propagation and stockout amplification in the context of supply chain inventory systems. Stockout amplification is a stage-to-stage increase in overall stockout rates. Stockout propagation is the tendency for stockout at one node to instigate a stockout at a neighbouring node and is conceptually related to the idea of cascading failures in physical systems, such as electrical power grids. We study these concepts in both upstream (‘supply side’) and downstream (‘demand side’) directions in the context of normal operating conditions for an adaptive R, S (periodic, order-up-to) inventory policy. We build a simulation model of a 5-stage serial supply chain that experiences normally distributed customer demands and gamma distributed lead times. We find that stockout propagation exists, but contrary to conventional wisdom, it occurs in the upstream direction. There is little indication that stockout propagation is occurring to any significant degree in the downstream direction. We also find stockout amplification occurring in the upstream direction in scenarios where more aggressively adaptive inventory parameter updating is performed. We discuss implications of this work in the areas of supply chain inventory modelling, ordering decisions, safety stock determination, and the use of adaptive inventory policies.     

A graph-based model to measure structural redundancy for supply chain resilience

Globalisation and lean initiatives increase the vulnerabilities of the supply chains (SC), where disruptions in any plant in a supply chain network (SCN) can propagate throughout the whole SCN. Redundancy is part of the SC re-engineering to improve supply chain resilience (SCRES). This paper presents a conceptual model of an SCN using graph theory, considering the relationships between plants and materials. Based on the model, the structural redundancy of the SCN is measured, which is used to assess SCRES. This assessment approach focuses on the resilience of the SCN against disruptions. Case studies are discussed to illustrate the applicability of this model and show that increasing structural redundancy of the SCN improves SCRES against disruptions.     

Towards a better understanding of supply chain quality management practices

This paper reports the results of a comparative study of quality tools and methods adoption by operations and supply chain managers. A survey was administered to both types of managers in the Western United States. Performing a Kruskal Wallis analysis, we found support for the hypothesis that operations and supply chain managers approach quality management differently. We found that operations managers tend to manage supply chains through procedural methods such as ISO 9000 and supplier evaluation. Supply chain managers tend to be more collaborative, emphasising supplier development and complaint resolution. We found that both types of managers adopted on the job training, data analysis, supply chain management, customer relationship management, project management and surveys. This paper represents another step in defining the field of supply chain quality management.     

Revolutionizing supply chain management the theory of constraints way: a case study

This research describes in detail an application of theory of constraints (TOC) and its resulting benefits on the supply chain performance of India's largest lock manufacturing company over a period of seven years. Using TOC's thinking process, the core constraints that had limited the company's performance in the areas of production, distribution, supply group and projects were identified and eliminated. TOC's unique approach helped the company achieve a significant reduction in its finished goods, raw material and work-in-process inventories at various levels across the supply chain. The stock-outs and excess in the distribution system nearly disappeared. The existing lead times saw a drastic reduction while the availability of items increased to nearly 100% despite significant decrease in inventory levels in the supply chain. The inventory turns of the distributors and retailers more than tripled and their profitability increased significantly. The overall sales of the company grew nearly three times during the six years post TOC implementation. TOC's holistic approach helped the company to double its profits and improve its cash position during the Great Recession.     

A control engineering approach to the assessment of supply chain resilience

There is no consensus on the supply chain management definition of resilience. To aid in evaluating the dynamic behaviour of such systems we need to establish clearly elucidated performance criteria that encapsulate the attributes of resilience. A literature review establishes the latter as readiness, responsiveness and recovery. We also identify robustness as a necessary condition that would complement resilience. We find that the Integral of the Time Absolute Error (ITAE) is an appropriate control engineering measure of resilience when it is applied to inventory levels and shipment rates. We use the ITAE to evaluate an often used benchmark model of make-to-stock supply chains consisting of three decision parameters. We use both linear and nonlinear forms of the model in our evaluation. Our findings suggest that optimum solutions for resilience do not yield a system that is robust to uncertainties in lead-time. Hence supply chains will experience drastic changes in their resilience performance when lead-time changes.