确定多目标问题有效最小值的新方法

An Exact Penalty Method for Finding Minimum Criterion Values over the Efficient Set

源于支付矩阵的最小值常被认为是多目标线性规划(MOLP)问题中有效集上真实最小值的一个“合理”近似,实际两者之间的差距常常是相当大的,为此,提出了一种确定有效集上真实最小值的精确罚函数法,并给出了相应的数值实例。

Minimum criterion values from payoff matrix are often integrated into a number of Multiple Objective Linear Programming (MOLP) problems. These values are usually considered as reasonably accurate estimates of the minimum criterion values over the efficient set that characterize the range of the criterion values over this set. However, the discrepancy between the minimum from the payoff matrix and the minimum over the efficient set can often be large. In this paper, a new exact penalty method is developed to determine the minimum criterion values over the efficient set. A numerical example with three criteria is presented and solved using this method.