We consider the space-efficient implementation of greedy algorithms for several fundamental problems on intervals. We assume a random access machine model with read-only access to input stored in Θ(n) words of memory, augmented with a random access memory (workspace) of size Θ(s) bits, where lgn ≤ s ≤ n. Our implementations are based on the efficient realization of an abstract data structure that we call a temporal priority queue that supports extract-min and advancetime operations for a static collection of entities, each of which is active for some pre-specified interval of time. This realization is a generalization of the memory-adjustable navigation pile proposed by Asano et al. in studying time-space tradeoffs for sorting. Using temporal priority queues we are able to implement familiar greedy algorithms for the maximum independent set problem and a variety of dominating set problems on intervals, using O(m(lg (sk/m) + n/s)) time and Θ(s) bits of workspace, where k is the size of output and m = min(sk, n). Choosing s = Θ(n) this achieves O(n lg k) output-sensitive time complexity for the maximum independent set problem on intervals, previously realized using Ω(n) words of workspace.
CITATION STYLE
Saitoh, T., & Kirkpatrick, D. G. (2017). Space-efficient and output-sensitive implementations of greedy algorithms on intervals. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10167 LNCS, pp. 320–332). Springer Verlag. https://doi.org/10.1007/978-3-319-53925-6_25
Mendeley helps you to discover research relevant for your work.