Refine
Document Type
- Article (1)
- Diploma Thesis (1)
- Doctoral Thesis (1)
Has Fulltext
- yes (3)
Is part of the Bibliography
- no (3)
Keywords
- contraction method (2)
- Analyse von Algorithmen (1)
- Banach spaces (1)
- Binärsuchbaum (1)
- Donkers theorem (1)
- FIND algorithm (1)
- Gaussian process (1)
- Kontraktionsmethode (1)
- Martingal (1)
- Quickselect (1)
Institute
- Mathematik (3)
We consider versions of the FIND algorithm where the pivot element used is the median of a subset chosen uniformly at random from the data. For the median selection we assume that subsamples of size asymptotic to c⋅nα are chosen, where 0<α≤12, c>0 and n is the size of the data set to be split. We consider the complexity of FIND as a process in the rank to be selected and measured by the number of key comparisons required. After normalization we show weak convergence of the complexity to a centered Gaussian process as n→∞, which depends on α. The proof relies on a contraction argument for probability distributions on càdlàg functions. We also identify the covariance function of the Gaussian limit process and discuss path and tail properties.
Within the last twenty years, the contraction method has turned out to be a fruitful approach to distributional convergence of sequences of random variables which obey additive recurrences. It was mainly invented for applications in the real-valued framework; however, in recent years, more complex state spaces such as Hilbert spaces have been under consideration. Based upon the family of Zolotarev metrics which were introduced in the late seventies, we develop the method in the context of Banach spaces and work it out in detail in the case of continuous resp. cadlag functions on the unit interval. We formulate sufficient conditions for both the sequence under consideration and its possible limit which satisfies a stochastic fixed-point equation, that allow to deduce functional limit theorems in applications. As a first application we present a new and considerably short proof of the classical invariance principle due to Donsker. It is based on a recursive decomposition. Moreover, we apply the method in the analysis of the complexity of partial match queries in two-dimensional search trees such as quadtrees and 2-d trees. These important data structures have been under heavy investigation since their invention in the seventies. Our results give answers to problems that have been left open in the pioneering work of Flajolet et al. in the eighties and nineties. We expect that the functional contraction method will significantly contribute to solutions for similar problems involving additive recursions in the following years.
Binärsuchbäume sind eine wichtige Datenstruktur, die in der Informatik vielfach Anwendung finden. Ihre Konstruktion ist deterministisch, zur Analyse ihrer Eigenschaften wird aber eine rein zufällige Eingabe zugrundegelegt. Viele Größe, wie z.B. Tiefe, Höhe und Pfadlänge werden seit Jahren viel untersucht. Als besonders interessant hat sich die Analyse des Profils, der Anzahl Knoten einer bestimmten Tiefe herausgestellt. In dieser Arbeit wird ein funktionaler Grenzwertsatz für das am Erwartungswert normierte Profil vorgestellt. Dazu werden unterschiedliche Zugänge gewählt, die hauptsächlich auf dem sogenannten Profil-Polynom beruhen. Zunächst wird ein klassischer Zugang mit Hilfe von Martingalen besprochen. Der diskrete Prozess wird dazu auf kanonische Weise in ein zeitstetiges Modell (Yule-Prozess) eingebettet. Ergebnisse im kontinuierlichen Prozess werden dann durch Stoppen auf den diskreten übertragen. Zudem wird ein neuerer Zugang vorgestellt, der auf der Kontraktionsmethode in Banachräumen unter Verwendung der Zolotarev-Metrik beruht.
