On The Influence Of Information In Ordering Behavior Of Inventory Management

Managers and researchers realize that traditional research on inventory management cannot be well applied in practice, because its mathematical models assume that decision makers are perfectly rational. It excludes behavioral factors that always exist during a decision making process. Accordingly, Behavioral Operations Management(BeOps) has attracted researchers’ attention. It focuses on empirical research on operating processes. By collecting real data, BeOps explores differences between human decisions and theoretical predictions, as well as the reasons for the differences. Based on empirical observations, BeOps supplements the theoretical research.Our research project employed laboratory experiment to study ordering behavior in contexts of censored newsvendor, competing newsvendors, and mutli-echelon supply chain system. The experimental results reveal that differences do exist between human ordering decisions and theoretical predicitons; the differences are induced by behavioral biases; the behavioral biases significantly depend on information configurations and profit parameters,while they can be well captured by specially designed behavioral models; more behavioral observations can be exposed by the behavioral models.Different from the behavioral literature on single newsvendor problem, we focused on a censored information scenario where the only available information was sales. A significant learning process and satisfactory learning results showed up by experimental data in the scenario. We designed an efficient learning model to trace decision makers’ learning pattern.Other than the censored information scenario, competing environment is also common in real market. In order to isolate the effect of competing information, we kept the ordering levels of single newsvendor problem and competing newsvendor problem the same by design. Comparing with the single newsvendor scenario, experimental data from the competing newsvendor scenario show that ordering levels increase in high profit groups but not in low profit groups; ordering variations increase in the low profit groups but not in the high profit groups. A behavioral model comprising mental accounting bias and quantal response equilibrium fits the ordering data well in the competing-newsvendor scenario.In addition, we considered a supply chain with five echelons, which is a typical inventory system in practice. By excluding all triggers of bullwhip effect proposed by literature, we aimed to compare the chains’ operating performances between no information sharing scenario and full information sharing scenario, and studied the value of sharing full information among all echelons within the chain. It turns out that the operating performances between the two information scenarios do not have significant distinctions. With linear regression models, we further discussed participants’ cognitions biases on inventory factors.The research project was conducted from an perspective of information. It studied the influence of information in three typical inventory problems. With the project, the experimental observations supplement the behavioral exploration in inventory management; the behavioral models provide alternatives of capturing real ordering behavior. It does not only feed managers with practical suggetions but also develops the inventory theory with behavioral evidences.