摘要: |
为了解决在数据不确定条件下极大熵权重配置模型找寻最优权重的问题,引入云模型将具有随机性和模糊性的数据转换为定性的数字特征。以此为基础在逼近理想解排序方法架构下定义了正负理想解,基于改进的欧氏距离设计了正负向距离,提出了多评估对象相对能力量化模型,并在极大熵准则下构建了新的权重配置模型,使得评估对象相对能力大小排序的专家知识得到充分运用。通过分析采用了模拟退火算法实现最优指标权重的搜寻,解决了原先算法模型不能应用于非凸优化情形的问题,进而给出科学合理的权重配置方案。仿真实例表明,在多次迭代后找寻到最优权重熵为1.377 7,满足要求,证明了所提算法的有效性。 |
关键词: 效能评估 云模型 逼近理想解排序 极大熵 权重配置 模拟退火 |
DOI:10.20079/j.issn.1001-893x.230212002 |
|
基金项目: |
|
A method of weight allocation for evaluation indicators under uncertain information |
SUN Wen |
(Southwest China Institute of Electronic Technology,Chengdu 610036,China) |
Abstract: |
In order to solve the problem that the maximum entropy weight allocation model finds the optimal weight under the condition of data uncertainty,the cloud model is introduced to convert random and fuzzy data into qualitative digital features.The positive and negative ideal solutions are defined under the Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS) method framework.Based on the improved Euclidean distance,the positive and negative distances are designed,the quantitative model of relative capability for multiple evaluation objects is proposed,and a new weight allocation model is constructed under the maximum entropy criterion which makes full use of the expert knowledge of ranking the relative ability of the evaluation objects.After analysis,simulated annealing algorithm is used to search the optimal index weight,thus solving the problem that the original algorithm model can not be applied to non-convex optimizetion.Then a scientific and reasonable weight allocation scheme is given.Simulation example shows,after several iterations,the optimal weight entropy is found to be 1.377 7,and the calculated optimal weight meets the requirements,which proves the effectiveness of the proposed algorithm. |
Key words: effectiveness evaluation cloud model TOPSIS maximum entropy weight allocation simulated annealing |