Median Selection Requires (2+epsilon)n Comparisons

Improving a long standing result of Bent and John [Proceedings of the 17th Annual ACM Symposium on Theory of Computing, Providence, RI, 1985, pp. 213--216], and extending a recent result of Dor, Hastad, Ulfberg, and Zwick [ SIAM J. Discrete Math., 14 (2001), pp. 299--311], we obtain a $(2{+}\epsilon)n$ lower bound (for some fixed $\epsilon>0$) on the number of comparisons required, in the worst case, for selecting the median of n elements.

[1]  Chee Yap.  New lower bounds for medians and related problems , 1976 .

[2]  David G. Kirkpatrick.  A Unified Lower Bound for Selection and Set Partitioning Problems , 1981, JACM.

[3]  Harold N. Gabow,et al.  A Counting Approach to Lower Bounds for Selection Problems , 1979, JACM.

[4]  Uri Zwick,et al.  Selecting the median , 1995, SODA '95.

[5]  Uri Zwick,et al.  Finding percentile elements , 1995, Proceedings Third Israel Symposium on the Theory of Computing and Systems.

[6]  Manuel Blum,et al.  Time Bounds for Selection , 1973, J. Comput. Syst. Sci..

[7]  Arnold Schönhage,et al.  Finding the Median , 1976, J. Comput. Syst. Sci..

[8]  Mike Paterson,et al.  Progress in Selection , 1996, SWAT.

[9]  David Galer Kirkpatrick,et al.  Topics in the complexity of combinatorial algorithms. , 1974 .

[10]  Ira Pohl,et al.  A sorting problem and its complexity , 1972, CACM.

[11]  Selmer M. Johnson,et al.  A Tournament Problem , 1959 .

[12]  J. Ian Munro,et al.  Average case selection , 1989, JACM.