Combining the inequalities (4.3)and (4.4)we obtain cost(GMS(I)-OptMs(I)≤pk≤ (4.4) (4.3) (/ =1 Finally,we bound the relative error by the following calculation: cost(GMS(I))-OptMs(I) (∑1)/k m ≤ OptMS(I) (4.5)】 (∑1p)/m (4.2) 到此为什么就可以说GMS算法是MS问题的2近似算法？到此为什么就可以说GMS算法是MS问题的2近似算法？