管理者有限理性决策行为对应急运作绩效的影响：模型与实验研究

As the case of Fukushima nuclear power plant damaged by tsunami in 2011, many researchers once again confirmed the factor that human are not entirely rational when making decisions. Lacking of enough information, needing to solve chaos in short time, fearing to be punished or loss reputations, managers could present to be much more bounded rational in decision rather than to be managed rationally for effectiveness. As a matter of factors, such kinds of decision behaviors could be frequently observed in disruption management of some vital important operation systems, such as power grid, communication and chemical plants, and managers seldom followed their pre-computed optimal plans or routines when critical capacities are crippled by unexpected events. We observed the phenomenon that there are always redundant capacities, such as flexible capacities, inventories and spare parts in upon large-scaled operation systems to avoid disruptions. And in case of suffering unexpected events, contingency plans are correspondingly started to cut the disruption cost. However, many evidences show that these pre-computed optimal reschedules and contingent plans are seldom followed or executed, especially in the situations of lacking enough information, needing to solve chaos in short time and fearing to be punished or loss reputation. All of these are the factors to make managers to be bounded rational. In order to find out what kinds of bounded rational behaviors will occur, and how much they will affect performance towards disruption management in Chinese scenario, we took carry an investigation on Zhejiang Electric Power Cooperation (ZEPC), and interestingly to find out 3 kinds of behaviors that could be notably influence managers' decision, they are loss aversion, ambiguity aversion/ seeking and reputation maintain/seeking. Based upon the findings, we intend to carry out our researches on disruption management through 2 kinds of methodologies, which are modeling and experiments. Firstly, we would like to use mathematical modeling methodology to create, expand or modify the now's models by further considering the 4 kinds of decision behaviors that we found in our investigation. And we hope to find out which kinds of parameters that would greatly influence the performance of disruption management, and contribute to the bias of theoretical model and reality. Through the mathematical modeling, we hope to give a much more explanations to the phenomenon that managers are always intend to over-use or under-use the flexible capacity and recovery. Secondly, we also present 4 experiments as well as computer simulations, for the purpose on one hand that to verify the mathematical models presented above, and on another hand find out the ways to promote performance of disruption management while suppress the negative and ineffective bounded rational behaviors. As the conclusion, we hope to add much more reasonable elements that could be helpful to enhance the explanation

应急情境下管理者的决策行为带有强烈的有限理性特点，忽视该因素会造成理论与现实之间存在较大的偏差。在前期调研的基础上（以浙江电力为对象），本项目明确了四类影响应急运作的管理者决策行为，并拟在电力、通信等大型运作系统能力因灾受损的情境下，以柔性能力扩张和受损能力恢复为抓手，以数理建模和实验两种方法展开研究。数理模型主要考虑了管理者损失厌恶行为、过度自信、模糊规避和声望寻求等常见的决策行为，并对现有应急运作模型进行拓展和修正，以期从不同视角对应急期间发生的应急不力和过度等现象的背后机理做出解释。实验研究则主要通过4个实验场和计算机模拟等，对上述管理者决策行为对应急绩效的影响进行分析，以期发现相关的行为控制机制。本项目将行为经济学和认知心理学等进行了融合交叉，拓展了行为运作管理在应急运作领域的研究，希望为现有研究增添更为合理的元素，以增强理论对现实的解释能力。

近年来发生的几个突发事件表明：在应急运作管理过程中管理者并非完全理性，有限理性的行为使得管理者在应急决策过程中产生了巨大的偏差，“应急过度”和“应急不力”是经常观察到的两个现象。以往应急运作管理的研究大多出于完全理性和风险中性假设进行模型的构建，而忽视了一个基本现实，即“人”这个在应急决策制定和执行过程中的表现是有限理性的，而这些有限理性行为如何造成应急决策偏差在应急运作管理领域中并没有很好地展开。基于此，本课题以电力、通信和石化等大型运作系统的应急为背景，研究了四种管理者有限理性决策行为对应急决策的偏差以及对应的纠偏机制。.经过有效分工与紧密合作，课题组完成标注的期刊论文8篇，完成或即将完成的工作论文2篇，专著初稿1部。其中，国家自然基金委认定的管理科学A类重要期刊论文3篇，EI检索论文2篇，SCI检索论文3篇，两篇工作论文分别与匹兹堡大学Katz商学院三位教授合作。本课题围绕五个模块展开研究，得到的主要结论有：一是，强制性行政监管机制下，存在一个明确监管惩罚下界，这既能迫使企业积极应急，同时又不至于恶化未来运营绩效。二是，当异常扰动造成需求集中爆发时，损失厌恶行为会完全改变管理者在完全理性下的能力结构投资方向，即投资更多的柔性能力。补偿投资同时施加高额的缺货惩罚，有助于减少柔性能力投资偏差。三是，过度自信行为是造成能力恢复不足的主要原因，“胡萝卜+大棒”的惩罚援助机制可以迫使管理者恢复更多的受损能力。四是，模糊效应会造成应急过度，管理者模糊信念在贝叶斯更新框架下能够快速降低。五是，声望寻求行为会促使管理者过度应急，政治声望是其中最重要的原因，高水平团队能够对政客型管理者的决策进行纠偏。.本课题以行为应急运作管理为主要研究对象，在研究过程中，融合了认知心理、行为经济和运作管理领域多种理论和研究方法，目的是在现有应急运作管理的理论和框架体系中增加管理者有限理性行为的元素，提升现有理论的解释能力，增强研究结果的适用性。