Modelling and Simulation of Purchasing Decision by Data-Driven Demand

This paper constructs a data-driven demand purchasing confidence rule-based model based on the theory of evidential reasoning to address the problem of purchasing decision bias in which the decision maker anchors the prior demand.First, we apply a synthetic algorithm to reason about the confidence structure distribution of the purchasing decision and then iteratively optimize the purchasing quantity according to the confidence structure distribution to minimize the cost loss function.Finally, we choose the average cost loss and service level value as the evaluation indexes to explore the purchasing confidence rule model.The simulation shows that compared with the expected purchasing decision based on the stochastic cumulative distribution function, the decision based on the confidence structure distribution can reduce the purchasing decision bias, manifested in the lower mean cost loss and a higher mean value of service level.Sensitivity analyses of the number of rules, product shelf life, and critical values show that the purchasing confidence rule base is highly adaptable.