Application of Epistemic Utility Theory to Satisfactory Optimization

Satisficing control remains an important concept in decision making. In this paper, a new epistemic utility satisficing control theory is proposed for a new model of complex CMMO (constrained multi-objective multi degree-of-freedom optimization) system. As well, an epistemic utility function is developed and used to adjust the feasible region of soft constraints. The theory proved in this paper indicates that the utility function not only expresses the subjectivity of the original satisfactory-degree function, but also takes the cost of searching for a solution into account. Thus, the satisfactory-degree function can be adjusted and its rationality can be validated. This theory contributes an analytical method to the inverse satisfactory optimization problem. The findings indicate that this theory has good convergence and outcomes desired for satisfactory-degree functions.     

基于下降搜索的量子进化算法

为了提高全局寻优能力和收敛速度，基于量子进化算法和混合遗传算法，提出了一种新的进化算法．该算法将下降搜索理论应用到量子进化算法中，改进了量子进化算法仅靠量子门进行迭代的作用，从而加快了收敛速度，并降低了个体在进化时产生退化的可能性．典型函数的仿真实验结果表明，该算法具有好的全局性和收敛性．