Abstract: To ensure the successful implementation of PPP projects, the public client and the private bidder must share risks of the project reasonably. This paper builds a set of suitable risk sharing scheme choosing index evaluation system for PPP projects, and designs a risk sharing scheme selection method based on GASA and rough set theory. GASA method is applied to deal with the problems of data’s low support after discretizing and influencing the accuracy of the decision making. Using the method based on rough set and conditional information entropy theory to calculate the weights of attributes in the index evaluation system improves the applicability and reliability and the accuracy of GRA method. It puts the weight and data after reduction into the GRA formula then takes the plan which has the moderate highest value as the optimal solution.

Key words: Public Private Partnership（PPP） project, risk sharing, rough set theory, Genetic Algorithm Simulated Annealing（GASA）, Grey Relational Analysis（GRA）

摘要： 为保证PPP项目的成功实施，公共部门和私营部门必须合理地分担风险。构建了一套适用于PPP项目的风险分担方案选择指标评审体系，设计了一种基于GASA和粗糙集理论的风险分担方案选择方法。引入GASA法解决数据离散化后对决策规则支持度低，影响决策准确性的问题。利用基于粗糙集和条件信息熵理论的方法计算权重，提高了适用性和可靠性，改良了GRA法的准确性。将求得权重和约简后数据代入GRA公式，选取适度值最高者为最优方案。

关键词: 公私合作制（PPP）项目, 风险分担, 粗糙集, 遗传模拟退火算法（GASA）, 灰色关联分析法（GRA）