相互关联R&D网络风险传播控制模型构建与仿真

Modeling and simulation of the control methods against risk propagation in interdependent R&D networks

考虑到风险传播可能会对研发网络造成严重后果,本文从界定风险负荷、定义风险容量、建立风险触发规则三方面,构建了相互关联R&D网络风险传播模型。在此基础上,从微观视角提出基于网络内部企业间风险分担机制的失效企业修复策略,从宏观视角提出通过外界力量投入修复已失效企业策略。进一步地,结合风险传播特征,提出了通过消除风险触发期望最大、风险触发期望最小、风险触发概率最大和风险触发概率最小四种失效边修复策略。仿真结果表明:微观视角与宏观视角下的风险控制策略均可有效提高研发网络面对风险传播时的鲁棒性;消除已发生风险集合中的风险个数越多,研发网络面对风险传播时的鲁棒性越高;四种修复策略中,风险触发期望最大与风险触发概率最大修复策略的控制效果最佳,其次为风险触发期望最小与触发概率最小修复策略。研究成果对于网络化背景下的研发风险管理具有一定借鉴意义。

With the globalization,networking,and informationization of the world economy,it is an inevitable trend for firms to form R&D collaboration with other firms in order to develop new products and gain competitive advantages.To cope with the rapid change of environment,firms usually participate in several R&D networks in a period of time.These projects are usually interdependent within a firm due to the limited resources,such as dependency on resources,technology,and equipment.Although R&D networks can bring a lot of advantages to firms,it is still a risky organization.When risks occur to a few firms in the network,it is possible to trigger the potential risk of other firms through cooperative relationships or project dependencies.As a result,the risk propagation phenomena might lead to the disintegration of the interdependent R&D networks.Therefore,it is of great significance to study methods of preventing and controlling risk propagation for improving the robustness of interdependent R&D networks.Firstly,considering quantifiable characteristics of risk and risk-resistance ability of firms,a firm will only“fail”when its risk load exceeds its risk capacity.Thus,the risk propagation phenomena in interdependent R&D networks can be regarded as the cascading failure in complex networks.Then,referring to load-capacity mode,a risk propagation model of interdependent R&D networks is constructed from three aspects,that is,defining risk load of firms,determining risk capacity of firms,and establishing the rules of risk propagation among firms.Based on the risk propagation model,two methods to repair the failed firms are proposed from the micro and macro perspectives.On the one hand,from the microscopic view,the method is proposed that the failed firms can recover by allocating excess risk load to these member firms.On the other hand,four risk eliminated strategies through external forces on networks are put forward considering risk interactions from the macro perspective.The strategies are eliminating risks with high expectation of triggering other risks,eliminating risks with low expectation of triggering other risks,eliminating risks with high probability of triggering other risks,eliminating risks with low probability of triggering other risks.Finally,numerical simulations are used to verify the effectiveness of risk control methods.The simulation results show that:(1)After adopting risk control methods,the robustness of R&D networks is greater than that of the R&D networks without adopting risk control measures,thus the effectiveness of risk control model is verified;(2)The more the number of risks eliminated in the set of risks,the higher the robustness of R&D networks in face of risk propagation and the better effect of risk control models are,which indicates that increase the number of risks eliminated can reduce the vulnerability of R&D networks against risk propagation to a certain extent;(3)Among the four risk elimination strategies,the robustness of R&D networks is the highest when adopting the two strategies of eliminating risk with high expectation of triggering other risks and with high probability of triggering other risks.Then,the risk control effect is followed by the strategies of eliminating risk with low expectation and low probability of triggering other risks.In addition,the risk control effect of eliminating risk with high expectation is close to that of high probability elimination strategy.The effect of the rest two risk control methods is similar.This paper proposes risk control models for interdependent R&D networks,and also verifies the effectiveness of the model.The research results provide a certain reference significance for firms and governments to prevent and control risk propagation under the background of networked R&D.