基于优化模拟退火算法的智能决策模型

Intelligent decision model based on optimization simulated annealing algorithm

决策理论在工业生产、管理决策、安全生产等越来越多的领域得到广泛应用,已经成为越来越多的研究者研究的重要课题。三支决策粗糙集模型作为一个重要的概率型粗糙集模型,在给定损失函数情况下可以导出多种概率型粗糙集模型,针对决策粗糙集模型构建的最优化问题,考虑到决策成本最小化,提出一个优化的模拟退火算法和启发式算法,从而得到代价最小的属性约简集,研究阐明了一种将粗粒度并行优化方法和启发式学习方法结合,解决粗糙集决策优化问题。实验证明提出的模拟退火的优化DTRS模型算法具有良好的有效性,运行时间也短于自适应算法,而且学习到的阈值能够得到较小的决策风险代价。研究揭示了优化表示带来的一些新的见解,对决策粗糙集模型的研究提供了新的思路。

Decision theory is widely used in industrial production,management decision-making,safety and other fields,which has become more and more hot reserch issues for researchers. Three-way decisiontheoretic rough set model is a probabilistic rough set model. It call derive several other probabilistic rough set models by setting corporate cost functions. Considering decision optimization of rough set constructing model and minimization cost of the decision,the research will give an optimized simulated annealing algorithm（ OSAA） and an adaptive algorithm for a minimum cost attribute reduction. A heuristic approach combined with the coarse-grained parallel algorithm are proposed to solve the problem of optimization DTRS. Experiments designed for the optimization simulated annealing algorithm and adaptive learning method are in comparing with the running time and decision-making cost. The optimization representation can bring some new insights into the research on decision-theoretic rough set model.