Building Optimum Reliable Systems through Value-at-Risk Criterion-based Two Stage Fuzzy Random Optimization Method

本文针对随机与模糊混合不确定性下的复杂决策问题，定义了基于VaR测度的模糊随机优化模型。在建立模型时，提出了综合启发式算法，并在VaR仿真中采用粒子群优化算法求解模型。所提出的模型使我们能够从风险的角度看待两阶段优化问题，解决混合不确定环境下目前难以解决的问题。此外，该算法能够近似计算VaR准则模糊随机系数。这种模型是独特的，作为设计和解决现实世界问题的两阶段VaR模型具有广泛的应用。

In this research, to deal with complicated decision making problems from a risk framework under hybrid uncertainty consisting randomness and fuzziness, Value-at-Risk (VaR) measure-based fuzzy random optimization models are defined. In building the models, synthesized heuristic algorithms are proposed and particle swarm optimization algorithm is used in VaR simulation to solve the models. The proposed models enables us to treat risk perspective of 2 stage optimization problems and solve the problems under hybrid uncertain environment, which are hard to solve till now. Furthermore, such proposed algorithms are capable of approximate computation of VaR criterion fuzzy random coefficients. Such proposed models are uniquely built with wide applications as two stage VaR models which design and solve real world problems.