Research On Robust Risk Programming Methods For Virtual Enterprise

Virtual Enterprise (VE) is a complex open business model that can help enterprises to respond rapidly to market demand, win in the fierce market competition. However, enterprises in a VE face more risks than a stand-alone enterprise. The problem of risk management of VE involves many factors, so it requires to be considered from the system point of view for these risk factors. Based on the characteristics of dynamic alliance itself, it is particularly important to establish an effective system of risk assessment for the virtual enterpriseThe cost for risk control measures is considered as a determine value in most of these researches. In fact, as the changing of time and conditions, the cost for risk control will change also. By considering the uncertainties, mean value method is usually used to eliminate the impact of uncertainty in some research. But this method is lack of robustness. When there is variability of parameters, the obtained feasible solution by means of mean value can not guarantee the feasibility or ensure very low probability of being feasible. Therefore, a new approach—Robust Optimization (RO) method is proposed here to deal with the impact of data uncertainty on the feasibility of solutions. When the number of variability parameters is less than the number of pre-specified, this approach can guarantee the feasibility of solution, it can also provide a probabilistic that guarantee robust feasible solution with high probability.In this research, for the characteristics of dynamic alliance and the fuzziness of its risk, it determines an effective evaluation method--fuzzy comprehensive evaluation method and HFMFs. To reduce the influences of the uncertainty of the cost for risk controls on virtual enterprise risk programming problem, Robust Optimization (RO) method is used to deal with the virtual enterprise risk programming problem. Three kinds of robust risk programming models are established while considering the uncertainty of the cost for risk controls, which are single strategy risk programming model, multi strategies single choice risk programming model and multi strategies multi choices risk programming model. ILOG optimization software is used to solve the problem. Finally, the numerical experiments suggested the effectiveness of the proposed method which eliminates the impact of uncertainty with a negligible increase in risk level.