Inventory Control in Supply Chain: a Model-Free Approach*

Inventory control in supply chains: Alternative approaches to a two-stage lot-sizing problem

In the real world, applications with very large state and action spaces and unknown state transition probability, classical reinforcement learning algorithms usually show poor performance. One way to address the performance problem is to approximate the policy or value function. Fuzzy rule-based systems are amongst the well-known function approximators. This paper presents a Flexible Fuzzy Reinforcement Learning algorithm, in which value function is approximated by a fuzzy rule-based system. The proposed algorithm has a separate module for tuning the structure of fuzzy rules. Moreover, the parameters of the system are tuned during the learning phase. Next, the proposed algorithm is applied to the problem of inventory control in supply chains. In this problem, a fuzzy agent (supplier) should determine the amount of orders for each retailer based on their utility for supplier, by considering its limited supply capacity. Finally, a simulation is performed to show the capability of the proposed algorithm.

The overall operations of supply chain are affected by the increase of demand uncertainty. Inventory control is an important part in supply chain management. Under the requirement mentioned above, this paper studies the inventory control with uncertainty demand. Three-echelon inventory model of manufacture regarded as core enterprise is modeled, which assumes that uncertainty demand subjects to normal distribution. Redundancy inventory in the node enterprise of supply chain is discussed under the two inventory replenishment strategy (continuous review inventory and periodic review inventory). It 's shown that the result of this study is valuable to reduce the total cost of inventory and improve the performance of supply chain.

A robust optimization approach to reduce the bullwhip effect of supply chains with vendor order placement lead time delays in an uncertain environment

Eastern route of south-to-north water transfer project is selected as a study case. Vendor management inventory (VMI) control theory is introduced into the study on water inventory control of the node lake of the project. When the possibility of the flood is taken into account, water inventory control models and water utilization profit models under independent control strategy, centralized control strategy and VMI control strategy are set up. Then VMI and independent control strategies of water resource are compared and restriction conditions of VMI control strategy are deduced. Finally, profit change curves of water buyers and water sellers under different inventory control strategies are obtained in simulation testing. The result shows that VMI control strategy can increase water utilization efficiency and improve the whole profits of water supply chain system. The study can provide guidance for the operations of south-to-north water transfer project and other inter-basin water transfer projects.

In contrast, there exist bullwhip effect in the supply chain of perishable product more easily. We establish a system dynamics model of the supply chain of perishable product, analyzing the causes of the bullwhip effects. Analysis shows that the changes of consumer demand, the differences of retailer's order cycle, wholesaler's order cycle, production cycle, the delivery delay of wholesaler and delivery delay of producer, retailer's stock, wholesaler's stock, producer's stock, etc, can interact, the effects are amplified step by step, finally bring bullwhip effects. By analyzing the conditions for convergence of the system, adjusting the order cycle and delivery delay time, optimizing the model, we can see that the response of the improved model tends to be a stabilized convergence function. Sharing sales data and inventory information, shortening the links of the supply chain, reducing lead time or delivery time, coordinating orders, establishing strategic partnership and cooperation mechanism of the enterprise strategy, etc, such measurement can eliminate bullwhip effects.

Multi-echelon inventory control is the main form of the management of supply chain inventory, and it has been paid extensive attention by scholars because of its importance in supply chain management. Process industries, represented by metallurgy, petrochemical and pharmaceutical industry, have typical characteristics such as high investment, high costs, and high resource utilization, they play a key role in industrialization. In this paper, the simulation model of a multiproduct three-echelon inventory control system is established for the hybrid supply chain of the process industry. Based on the idea of control engineering, a feedback control law is designed for inventory control. Several mixed inventory control strategies were applied to the model. The proportional plus integral control algorithm is used to modify the inventory control strategies. Based on this simulation model, a simulation-based optimization method is applied to the three-echelon inventory control system. The results show that the reciprocal of total entropy ratio and customer satisfaction are optimized by the mixed inventory control strategies when the hybrid supply chain is affected by order uncertainty.

Supply chain management (SCM) is defined as a set of approaches utilized to efficiently integrate suppliers, manufacturers, distributors and retailers to make sure that merchandise is produced and distributed at the right quantities, to the right locations, and at the right time, in order to minimize system-wide costs while satisfying customer requirements. As a long-term cooperation strategy among enterprises, supply chain requires stable operations as necessary conditions for optimization and long-term competitive advantage of the whole supply chain. Stability, therefore, becomes an important indicator in supply chain management. Research on supply chain stability began in early 1960s. Based on system dynamics method, Forresterう1961え proves the presence and the cause of the bullwhip effect. He indicates that bullwhip effect could be weakened by improving decision-making behavior. From then on, the bullwhip effect and supply chain stability have been studied by many researchers. Most of these studies employed an ordering policy known as the APIOBPCS model (John et al., 1994; Disney and Towill, 2003). APIOBPCS policy is optimal in that, the policy minimizes the variance of the inventory levels with a sequence of forecast errors of demand over the lead time given (Vassian, 1954). In discrete time domain, pure time delays are readily handled by the z-transform and many results are known about its stability (Disney and Towill, 2002), variance amplification properties (Disney and Towill, 2003; Disney et al., 2004), and dynamic performance (Dejonckheere et al., 2003). Disney et al. (2006) investigate the stability of a generalized OUT (Order Up To) policy for the step response in both discrete and continuous time. Disney et al. (2007) prove that discrete and continuous Bullwhip Effect expressions have similar structures and show that the two domains are managerially equivalent and each domain can be used to study a supply chain in practice. The pure time delay causes complications in the time domain differential/difference equations. In the frequency domain, such equations generate an infinite number of zeros in the transfer function. Certain progress has been made on the stability of such systems by recasting the policy as a Smith Predictor (Warburton et al., 2004; Riddalls and Bennett, 2002). However, little is known about variance amplification issues or other aspects of the model’s dynamic performance. The use of Lambert ω function to solve such problems has recently gained some popularity. Several authors have studied the production and inventory control problem using continuous mathematics; and in the solution process the Lambert ω function is used (Warburton, 2004a, b; Warburton et al., 2004).

The existence of time delay causes the operation of supply chain system to aggravate the instability, induce bullwhip effect and reduce overall performance. In this paper, a discrete mathematical model with time-varying delay is constructed to investigate supply chain inventory network system with time-varying delay. Based on the proposed discrete mathematical model with time-varying delay, the inventory control strategy is desgined by using Lyapunov-Krosovskii functional analysis method. The feasibility analysis of supply chain inventory control strategy on bullwhip effect and the influence factors of time-varying delay are discussed by means of numerical simulation. The results prove that the inventory control strategy can restrain bullwhip effect effectively. It provides a novel approach to study supply chain inventory network system with time-varying delay from discrete perspective.

Introduction to distribution logistics

The competition in the twenty-first century is not the competition between enterprises, but the competition in the supply chain. “Bullwhip effect” is a common phenomenon in supply chain management. With the development of economy and the intensification of competition, the structure of supply chain is increasingly complex. Therefore, how to eliminate and effectively prevent the harm of “Bullwhip effect” has become a problem that people pay special attention to. The objective of this study is to provide a new way of using agility to prevent bullwhip effect. Through the combination and application of a variety of analytical methods, the key criteria of agility are ranked, so that enterprises can control these factors in priority, and formulate strategies in advance, so as to improve the agility of enterprises to prevent the negative impact brought by the bullwhip effect, and find the optimal solution to the “bullwhip effect.” Taking a Guangdong Yinglian Technology Company as an example, this study found that the best way to reduce the “bullwhip effect” of the supply chain is to improve customer satisfaction, and secondly to improve employees’ work capabilities.KeywordsAgilityBullwhip effectQFD

The supply chain design plays a key role in product availability in normal operations. The process of availability of stock level in running a business is necessary to attract and retain customers. The components or the channel partners in supply chain are responsible for calculating forecasted demand for the products. The overestimation as well as underestimation of the forecasted demand by each channel partners in the supply chain may result in the phenomenon of bullwhip effect in the supply chain. Vegetable, fish, and meat supply chains, which are the highly perishable products of daily life use, are to some extent different from that of the non-perishable products’ supply chains. Thus, the channel partners involved in the supply chain of vegetable, fish, and meat also forecast the different stock levels of the respective categories. In some cases, there are excess or shortfall in the availability of such items. This study attempts to comprehend and emphasize the phenomenon of the bullwhip effect in the category of fish and vegetable supply chain. The study further assesses the underlying factors responsible for the bullwhip effect. The study concludes with major suggestions for eliminating such bullwhip effect.

Cost-Effective Production and Inventory Control in Multi-echelon Supply Chains with Customer Prioritization

PurposeThe operational aspects of supply chain, when handled correctly, results in diminishing the impact of the bullwhip effect. The purpose of this study is to analyze the impact of operational and financial variables on the bullwhip effect. Various operational factors that contribute to the bullwhip effect in a supply chain are identified and their impact on variability in production is measured at manufacturer’s end in the supply chain.Design/methodology/approachTen different sectors of the Indian economy are identified and analyzed on the basis of bullwhip effect. The ratio of change in production with respect to change in demand is taken as a metric to measure the bullwhip effect. Initially, the impact of identified variables on bullwhip effect is analyzed using the linear regression analysis and then to gain more insights, the threshold regression model is applied according to the change in bullwhip ratio.FindingsThe study identifies four threshold regions in which bullwhip ratio is changing its slope considerably. The operational and financial variables impacting bullwhip effect differently in these four regions provide useful insights about how the variables are impacting the bullwhip effect.Research limitations/implicationsPast 11 years of observations on identified operational and financial variables are studied for ten different sectors. The operational and financial variables are identified on basis of available literature but may not be exhaustive in nature.Practical implicationsThe present study implies that the emphasis must be given to the magnitude of the bullwhip ratio. Strategies must be adopted that result in mitigation of bullwhip effect. Such mitigation strategies must not only be restricted on the basis of type of product or sector, perhaps they must be on the basis of threshold region of bullwhip ratio.Originality/valueThe study suggests a novel approach to study the bullwhip effect in supply chain management using the application of threshold regression considering the bullwhip ratio as a threshold variable.

Symbiotic Supply Chains – where the inputs of some members belonging to one supply chain are obtained using the waste generated by other companies in a different supply chain – represent a great opportunity for companies to move towards the Circular Economy paradigm. However, the dynamic behavior of Symbiotic Supply Chains (i.e., how detrimental time-varying phenomena of supply chains, such as the bullwhip effect, impact the performance) remains largely unknown. In this work we aim to contribute to this quite unexplored research field by analyzing the dynamics of two supply chains that implement, at the manufacturer level, a symbiotic exchange of waste. We derive closed analytical equations to estimate several indicators of the bullwhip effect under different information sharing scenarios and evaluate them for different coefficients of circularity (i.e. the average proportion of the output of one supply chain that is obtained using the waste of the other supply chain). Furthermore, we extend our analysis by carrying out a simulation study, when some underlying assumptions are removed, and a full-factorial Design of Experiments to emulate and explore the behavior of Symbiotic Supply Chains under further real-life scenarios. By doing so, we identify different properties of Symbiotic Supply Chains and provide some managerial implications. Among those, we observe how the bullwhip effect of one supply chain propagates to the other, which is also greatly influenced by the variability of the demand in both supply chains. Also, we note that high coefficient of circularity does not necessary improve the dynamic performance of Symbiotic Supply Chains. Results point out to the need to carefully design symbiotic exchange in supply chains to avoid undesirable effects that may hinder the advocated advantages of Circular Economy practices.