Font Size: a A A

Researching The Newsvendor Decision Bias Based On Bounded Rationality From The Perspective Of Behavioral Experiments

Posted on:2016-02-27Degree:DoctorType:Dissertation
Country:ChinaCandidate:X D DingFull Text:PDF
GTID:1319330512961145Subject:Logistics Engineering
Abstract/Summary:PDF Full Text Request
The classical newsvendor model assumes that decision maker is risk-neutral and aims to maximize its expected profit. However, faced with the random market, decision maker's actual order behavior is affected with psychological factors such as risk preferences, management experience, cognitive bias and so on, resulting the bias between the actual order behavior and the newsvendor model. Newsvendor decision-making bias is one of the typical bias. Firstly, this paper researched decision-maker's perception to stochastic market, then, studied the formation mechanism of newsvendor bias from the perspective of waste/ stock-out aversion, loss aversion and risk aversion. Finally, this paper constructed individual decision model based on the "anchor-adjustment" heuristic theory and group decision model based on the prospect theory.(1) Effects of stochastic market on decision-maker's order behavior?.This paper researched decision maker's perceptual behavior to history stochastic demand. Experimental data indicates that decision-maker is very significant to the last and average demands. Based on the anchor adjustment models, this paper built the group and individual models respectively and used them to explain the newsvendor bias phenomenon. When decision-maker does not focus on the historical demand data, there is no significant difference among the order behaviors. While when the decision-maker concerns the historical demand data, there is significant difference between the order behaviors. Mean anchor model can not describe the actual ordering behavior. Such decision maker's order behavior is dynamic and complex. We can use individual perceive model to predict such order behavior.?.This paper researched decision maker's perceptual behavior to uncertain demand. Based on the classic newsvendor model, this paper introduced probability choice theory and found that the perceived probability of demand for decision maker does not meet the bayesian axioms, but the truncated normal distribution. Then this paper proposed the optimal inventory decisions considered perception utility and measured the decision noise of different products by behavioral experiments. Combining the experimental data, one simulation experiment was done to simulate decision maker's order behavior under different cost. Results show that the perception utility model can explain the pull to center, PTC, phenomenon within the inventory decision-making. What's more, when the production costs c approaches (p+s) / 2, the PTC phenomenon will gradually disappear. At the same time, the unit PTC degree will decrease with the increasing of c. Thus it can be seen that PTC does not one universal phenomenon, but exists in the order behavior of some products only.(2) The impact of subjective bias avoidance to order behavior?.This paper researched the influence of waste/stock-out aversion bias on order behavior. Based on the classic newsvendor model, this paper introduced the waste-averse and the stockout-averse preferences to research decision maker's order behavior, finding that order behavior is impacted by the waste-averse and stockout-averse preferences. The order quantity is positive with the stockout-averse preference and negative with the waste-averse preference. Simulation experiments reveal that decision maker anchores on the optimal solution of the newsvendor model firstly and then adjusts to the direction of mean demand. The degree of adjustment is impacted by the coefficient of waste-averse and stockout-averse preferences, and the adjustment behavioral meets the characteristics of prospect theory value function curve.?.This paper researched the influence of the loss aversion bias on order behavior. This paper introduced stckout averse factor to modify the classic newsvendor model, and construct newsvendor model with loss averse under stockout punishing. According to the new model, we learn that stockout penalty coefficient decreases with product profit. Decision maker exhibits different order behavior among different products. In high-profit(low-profit) product, decision maker with loss reverse would order more(less) than the risk neutral decision maker, while the order quantites are equal within the balance-profit product. When decision maker is affected by both stockout averse and loss averse, he behaves newsvendor deviation. Finally, we take behavior experiment method to verify the above conclusions. Results show that both the individual and group order behavior could get through hypothesis test.?.This paper researched the influence of the risk aversion bias on order behavior. Based on the CARA utility function, this paper explained the formation mechanism of newsvendor decision bias from the perspective of risk preferences. The results of behavioral experiments show that decision maker's risk preferences will reverse with relative cost increase. In addition, decision maker exhibits risk aversion (love) order behavior under the high (low)-profit market. What's more, risk aversion has more significant effect on decision maker's order behavior.(3) Establishing the order behavior model based on bounded rationality deviation?.Experiments of risk appetite and inventory decision-making behavior show that newsvendor bias can not explain the phenomenon of individual behavior. The influences of the last demand and risk preferences are very significant for decision-maker's inventory behavior. This paper proposed the psychological desired point based on the individual risk preferences, and estimated the psychological desired point through the regression model. Concerned with the last demand and the psychological desired point, this paper constructed one anchor-adjustment inventory decision behavior model. This paper proposed that decision-maker orchors on the last demand and adjusts t towards to the psychological desired point. The adjustment is affected by the risk preference and decision-making scene mainly. The order quantity of risk lover is significantly higher than risk aversion and risk neutral. Decision maker exhibits aggressive-confident order behavior in the follow-decision scene, and exhibits avoidance-hesitant order behavior in back-decision scene.?.Prospect theory is the basic theory in the field of uncertain decision making. Nagarajan thought that prospect theory could not explain the newsvendor decision-making bias when the reference point is zero-profit. Recently, Long thought that under the mean prospect model, the prospect theory could explain it. Therefore, through the improvement of mean prospect, this paper instructed the computing method of reference point for different products, and put forward the new inventory decision-making model under the prospect theory. MATLAB simulation results exhibit that in different products, decision maker's psychological reference point and sensory loss (benefit) interval is significantly different. What's more, in part of products, the prospect theory could explain decision maker's actual order behavior. This conclusion is consistent with Nagarajan and Long, and theoretically proves the relationship between the prospect theory and the newsvender decision-making deviation.This paper studied decision maker's limited rationality deviation that may exhibit under the stochastic demand and analyzed the impact of bounded rationality deviation on order behavior. Then analyzed the relationship between the averse bias and newsvender decision bias by theoretical derivation. Finally, behavioural experiments were used to validate the model, and relevant strategy to remove decision bias was proposed.
Keywords/Search Tags:Behavioral operation management, Inventory decision-making, Newsvendor model, Newsvendor decision-making biases, Prospect theory, Behavioral experiment
PDF Full Text Request
Related items