log-Lists and their applications to sorting by transpositions, reversals and block-interchanges

• Published in: Theoretical computer science Vol. 660; pp. 1 - 15
• Main Author: Rusu, Irena
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 17.01.2017; Elsevier
• Subjects: Computer Science; Data Structures and Algorithms; Double-linked list; Dynamic list ranking; Link-cut tree; Permutation sorting; Permutation sorting; Link-cut tree; Double-linked list; Dynamic list ranking
• ISSN: 0304-3975; 1879-2294
• DOI: 10.1016/j.tcs.2016.11.012

Link-cut trees have been introduced by D.D. Sleator and R.E. Tarjan [12] with the aim of efficiently maintaining a forest of vertex-disjoint dynamic rooted trees under cut and link operations. These operations respectively disconnect a subtree from a tree, and join two trees by an edge. Additionally, link-cut trees allow to change the root of a tree and to perform a number of updates and queries on cost values defined on the arcs of the trees. All these operations are performed in O(log⁡n) amortized or worst-case time, depending on the implementation, where n is the total size of the forest. In this paper, we show that a list of elements implemented using link-cut trees (we call it a log-list) allows us to obtain a common running time of O(log⁡n) for the classical operations on lists, but also for some other essential operations that usually take linear time on lists. Such operations require to find the minimum/maximum element in a sublist defined by its endpoints, the position of a given element in the list or the element placed at a given position in the list; or they require to add a value a, or to multiply by −1, all the elements in a sublist. Furthermore, we use log-lists to implement several existing algorithms for sorting permutations by transpositions and/or reversals and/or block-interchanges, and obtain O(nlog⁡n) running time for all of them. In this way, the running time of several algorithms is improved, whereas in other cases our algorithms perform as well as the best existing implementations.