A performance guarantee for the greedy set-partitioning algorithm
暂无分享,去创建一个
SummaryLet S be a set of positive numbers and m an integer not less than 2. The problem is to partition S into m subsets such that the ratio of the largest subset sum to the smallest is as small as possible. Let ϱg(S) be the value of this ratio using the greedy or largest-first rule and ϱ0(S) be the smallest possible value of this ratio, i.e., the optimal value. The authors prove that
$$\frac{{\rho _g \left( S \right)}}{{\rho _0 \left( S \right)}}\underset{\raise0.3em\hbox{$\smash{\scriptscriptstyle-}$}}{ \leqslant } {7 \mathord{\left/ {\vphantom {7 5}} \right. \kern-\nulldelimiterspace} 5}$$
, and that this is a best possible bound for all m.
[1] D. K. Friesen,et al. SCHEDULING TO MAXIMIZE THE MINIMUM PROCESSOR FINISH TIME IN A MULTIPROCESSOR SYSTEM , 1982 .
[2] Edward G. Coffman,et al. A generalized bound on LPT sequencing , 1976, SIGMETRICS '76.
[3] Ronald L. Graham,et al. Bounds on Multiprocessing Timing Anomalies , 1969, SIAM Journal of Applied Mathematics.
[4] Richard M. Karp,et al. The Differencing Method of Set Partitioning , 1983 .