A Nearly Optimal Algorithm for Approximate Minimum Selection with Unreliable Comparisons
We consider the approximate minimum selection problem in presence of independent random comparison faults. This problem asks to select one of the smallest k elements in a linearly-ordered collection of n elements by only performing unreliable pairwise comparisons: whenever two elements are compared, there is constant probability that the wrong answer is returned. We design a randomized algorithm that solves this problem with high probability (w.h.p.) for the whole range of values of k using O( n · ( n/k + n )) expected time. Then, we prove that the expected running time of any algorithm that succeeds w.h.p. must be Ω(n/k n), thus implying that our algorithm is optimal, in expectation, for almost all values of k (and it is optimal up to triple- factors for k = ω(n/ n)). These results are quite surprising in the sense that for k between Ω( n) and c · n, for any constant c<1, the expected running time must still be Ω(n/k n) even in absence of comparison faults. Informally speaking, we show how to deal with comparison errors without any substantial complexity penalty w.r.t. the fault-free case. Moreover, we prove that as soon as k = O( n/ n), it is possible to achieve a worst-case running time of O(n/k n).
READ FULL TEXT