最佳协调解的分步筛选原理及应用

ASIEVE PRINCIPLE OF THE BEST COMPROMISE SOLUTION AND APPLICATION

本文针对研究大系统问题建立的多目标决策模型的特点,提出使用分步筛选方法求其最佳协调解。首先,通过使所求目标值与最优目标值的允许误差来反映决策人的偏好;其次在满足决策人偏好的条件下,使用优序数原理选择能使决策模型达到总体最佳的满意解为该多目标决策问题的最佳协调解。证明了所选最佳协调解必为非劣解,并给出了案例。

In this paper a sieve method based on the characteristic of multiobjective decision model researched large system problem is proposed to find the best compromise solution. First, decider's preference is considered by the permissible error that objective value obtained compares with the best goal value. Under the condition that the decider's preference is satisfied, the best compromise solution is found by the principle of optimal coefficient. We prove the best compromise solution is a noninferior solution. At last, a case study is given.