R&D项目组合鲁棒性风险测度及选择模型研究

Research on Measurement of Robustness Risk and Selection Model of R&D Project Portfolio

针对R&D项目组合选择研究中将风险作为极值目标函数,导致所求最优函数解为折衷解,且研究中较少考虑项目依赖关系对风险变化及组合选择影响的问题,借鉴鲁棒性理论,提出项目组合鲁棒性风险概念,并从单项目风险、项目依赖关系和项目风险控制成本三方面描述其构成。采用相继故障理论对项目组合鲁棒性风险进行刻画,建立了以鲁棒性风险、项目资源为约束,收益最大化为目标函数的组合选择优化模型,并采用量子遗传算法进行求解。实例分析表明,最小化风险组合选择模型排除了一些进入组合后能承担自身风险且增加收益的项目,而考虑项目组合鲁棒性风险的选择模型不但能为项目决策者有效地决策支持,而且能降低因选择不合理而导致的项目组合失败的可能性。

Facing the problems of R＆D project portfolio selection of which most of literatures regard risk as the minimizing object function, leading to the optimal solution converting into a compromise solution, and consider- ing few studies that have been made on the influence of interdependency relationship on the change of project risk and the project portfolio selection, the conception of project portfolio robustness risk is proposed based on robust- ness theory. The definition is described from three aspects, that is, project risk, interdependency relationship and control cost for project risk. Then project portfolio robustness risk is measured with the theory of cascading failure and a project portfolio optimization model is established to maximize the profit object and satisfy the pro- ject portfolio robustness risk and resource constraints. To solve this model and illustrate its effectiveness, the quantum genetic algorithm is employed and an example is given. The result shows that the risk minimization model excludes these projects that can increase profits and take risks after they are selected into the portfolio. However, the model developed in this paper can not only provide effective decision support for the decision maker but also decrease the failure possibility of projects with the unreasonable selection.