The overconfident and ambiguity-averse newsvendor problem in perishable product decision. (October 2020)

• Record Type: Journal Article
• Authors: Wu, Dasheng; Chen, Feng
• Is Part Of: Computers & industrial engineering. Volume 148(2020)
• Publication Date: 2020-10
• Subjects: Newsvendor problem -- Ambiguity-averse -- Overconfidence; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2020.106689
• Languages: English
• ISSNs: 0360-8352

Abstract

Abstract: In the decision-making of perishable product, newsvendors may be overconfident as they have a more precise estimate about future outcomes, and they may be ambiguity-averse as only partial knowledge of the demand distribution is available. This paper investigates the overconfident behavior bias of a newsvendor when he is ambiguity-averse. By reducing the variance of demand distribution but preserving the mean with a linear function, we formulate a theoretical max–min optimization model that accounts for the influence of overconfidence and ambiguity aversion on the newsvendor's decision. The closed-form solution for optimal order quantity is achieved by maximizing the expected profit against the worst distribution. Some conclusions are accordingly included. Firstly, the optimal order quantity is a linear function of the overconfident level, and it is decreasing in the overconfident level in the high-profit case, and increasing in the low-profit case. Secondly, overconfident newsvendors believe they are more precise, and thus gain more profits than the unbiased ones. However, overconfidence may actually raise order bias, and the expected costs increase in the overconfident level. Thirdly, overconfidence helps to alleviate the conservatism of ambiguity-averse optimization and the robust strategy performs reasonably better even if the demand distribution exists some misspecifications. In addition, we provide experiments supporting the theoretical relations. The closed-form solutions are tractable, and thus of great significance to the practical applications. Highlights: Closed-form solution for ambiguity-averse and overconfident newsvendor problem. Optimal order quantity is a linear function of its overconfidence level. In the ambiguity-averse setting, the overconfident newsvendor shows order bias. The expected cost of order bias increases by the level of overconfidence. Overconfidence helps to alleviate the conservatism of ambiguity-averse optimization.