关键词: 泊松过程和Marshall-Olkin分布, 多重风险流, CVaR准则, 多级供应链决策模型

Abstract: Under the complex background of great changes of the century, the epidemic of the century and the conflict between Russia and Ukraine, the uncertainty of supply chain has the essential characteristic of diversification, and the multiple risk flows resulting from the evolution bring great challenges to the theories and methods of supply chain operation and management. For example, in the decision-making environment of multiple risk flow impact, the following series of problems remain to be further solved and studied: First, how do we characterize the transmission mechanism of multiple risk flow in each node of the supply chain? Second, how do we incorporate the statistical regularity of multiple risk flows into the theoretical framework of the decision model? Third, does the risk aversion factor held by the manager have only one dimensional attribute to the mechanism of decision making? The existing research results in the academic circle support the view that the action mechanism of risk aversion factors is a single image attribute, that is, the higher the degree of risk aversion, the more conservative the decision-making behavior. Therefore, guided by the above problems, this paper proposes a new decision model based on Poisson process and Marshall-Olkin distribution, combined with the basic theory of the newsboy model, and then reveals the action mechanism of the impact degree of multiple risk flows on the operation and management mechanism of the supply chain, with a view to expanding the universality of supply chain decision theory and method in practical application.
In the context of multiple risk impacts, combined with the theoretical results of this paper and the experimental results of the above numerical simulation, the transmission mechanism of risk aversion factors on supply chain operation and management can be fully revealed, namely:
   First, there is a negative correlation between the derived expected profit and the degree of risk aversion in type I (demand randomness——capacity determination), type II (demand randomness——capacity randomness), type III (demand randomness——capacity randomness) and other cases, that is, when the information factors carried by multiple risk flows increase the degree of risk aversion, the expected profit of supply chain tends to decline.
Second, under the impact of multiple risk flows, the optimal order quantity derived from type I (demand random——capacity determination) and type II (demand random——capacity randomization) is a monotonically decreasing function of risk aversion. This conclusion is consistent with the existing academic views, but it has not been able to further promote the academic understanding of the transmission mechanism of risk aversion factors.
Third, under the impact of multiple risk flows, the optimal order quantity derived from the condition of type III (demand determination——capacity randomization) is a monotonically increasing function of risk aversion. Obviously, this conclusion is the opposite of the conclusion obtained under Type I and Type II. Therefore, it shows that risk aversion factors have the property of “two opposites” to the transmission mechanism of supply chain, that is, under type I and type II, the higher the degree of aversion of decision makers, the more conservative the ordering strategy they adopt. Under type III, the higher the degree of risk aversion held by decision-makers, the more aggressive the ordering strategy they adopt, with a “sister madness” type decision-making tendency. It can be seen that the transmission mechanism of risk aversion factors in traditional academic circles is confined to the cognition that “the higher the degree of risk aversion, the more conservative the decision-maker”, which has limitations in dimensions and horizons. Therefore, the conclusions derived under Type III are helpful for the academic community to deepen the understanding of the transmission mechanism of risk aversion factors, and further facilitate the research achievements in the field of risk decision-making to make breakthroughs.

Key words: Poisson process and Marshall-Olkin distribution, multiple risk streams, CVaR criterion, multi-level supply chain decision model