本文基于某电力公司的实地调研,在运作能力因灾受损情境下研究了应急修复过程中管理者两种过度自信行为对决策偏差的影响,分析了惩罚援助机制在决策纠偏方面的作用.研究结果表明:提高惩罚和积极援助的“胡萝卜+大棒”调控机制有助于纠正管理者过度自信造成的决策偏差;过度精确行为对偏差存在双向影响,但可通过调节惩罚援助参数比例促使管理者向“积极应急”的方向决策;过高估计行为必将造成应急投入不足,同时将抑制过度精确行为的偏差,惩罚援助的调控机制必然失效;外界随机扰动分布左偏时,管理者需适当增加决策量以缓解应急投入不足.%Overconfidence behavior stems from the confidence of manager based on his prior knowledge,which could be strengthened by learning,training or successful experience.Some theoretical and experimental studies suggest that overconfidence could help a manager improve his decision-making performance in the scenario of having incomplete or uncertain information.However,this is not always the case,especially in the scenario of unexpected events with extremely low probability,such as fire,earthquake,and terrorist attacks.Decision bias caused by overconfidence behavior could weaken or even invalidate a manager's performance,because the kind of human behavior might make manager lose his capability of knowledge calibration when he observes the inconsistencies with his prior knowledge.Some cases,such as the Blackout in USA and Canada(2003),and Fukushima nuclear leak in Japan(2011),verify that a manager's overconfidence plays a seriously negative role during the disruption events,which is also revealed by our survey of a power company's senior and middle managers.The impact of a manager's overconfidence towards the disruption management'performance is seldom studied in the field of disruption management.Moreover,researching a mathematical model that includes different kinds of overconfidence is even less in the voluminous operational researches.This paper examines the disruptive situation of large-scale operation systems,such as power station,chemical plant and communication systems,when their critical capabilities are crippled by unexpected events.The goal is to understand how decision bias causes a manager's overconfidence,and further analyze the impact of this kind of behavior towards regulatory penalty and subsidy mechanism.In section 2,we firstly construct a mathematical model when manager is assumed to be completely rational.This model is called Model I.This model includes all kinds of cost that could be met in the period of disruption management,and it is proved to be a newsvendor model.Model Ⅰ serves as a basic model for subsequent comparison.In section 3,we extend the Model Ⅰ to Model Ⅱ by including two kinds of overconfidence behavior which are over-estimation and over-precision.We prove that a manager's decision is inevitably biased in Model Ⅱ when compared to Model Ⅰ,and decision bias caused by these two kinds of overconfidence behavior is rather different,which also affects the calibration capacity of regulatory penalty and subsidy.In section 4,we present the numerical analysis in order to show a manager's decision bias when random disturbance is symmetrically and asymmetrically distributed.We find that over-estimation and over-precision are consistent behaviors in the process of recovering injured capacities.However,bilateral effect on decision bias would enlarge or narrow the bias gap when manager shows over-precision.In section 5,we present 5 managerial insights in order to show the ways for both regulators and managers to improve their decision-making performance.For regulators,4 managerial insights could help narrow the gap of the decision bias when a manager is overconfident.As for an overconfident manager,one important insight is helpful to lower the negative impaction during disruption management.
展开▼
