Search results for "OPTIMIZATION"
showing 10 items of 2824 documents
Asymptotic optimality of myopic information-based strategies for Bayesian adaptive estimation
2016
This paper presents a general asymptotic theory of sequential Bayesian estimation giving results for the strongest, almost sure convergence. We show that under certain smoothness conditions on the probability model, the greedy information gain maximization algorithm for adaptive Bayesian estimation is asymptotically optimal in the sense that the determinant of the posterior covariance in a certain neighborhood of the true parameter value is asymptotically minimal. Using this result, we also obtain an asymptotic expression for the posterior entropy based on a novel definition of almost sure convergence on "most trials" (meaning that the convergence holds on a fraction of trials that converge…
Hölder Continuity up to the Boundary of Minimizers for Some Integral Functionals with Degenerate Integrands
2007
We study qualitative properties of minimizers for a class of integral functionals, defined in a weighted space. In particular we obtain Hölder regularity up to the boundary for the minimizers of an integral functional of high order by using an interior local regularity result and a modified Moser method with special test function.
A fast and recursive algorithm for clustering large datasets with k-medians
2012
Clustering with fast algorithms large samples of high dimensional data is an important challenge in computational statistics. Borrowing ideas from MacQueen (1967) who introduced a sequential version of the $k$-means algorithm, a new class of recursive stochastic gradient algorithms designed for the $k$-medians loss criterion is proposed. By their recursive nature, these algorithms are very fast and are well adapted to deal with large samples of data that are allowed to arrive sequentially. It is proved that the stochastic gradient algorithm converges almost surely to the set of stationary points of the underlying loss criterion. A particular attention is paid to the averaged versions, which…
On-line Construction of Two-Dimensional Suffix Trees
1999
AbstractWe say that a data structure is builton-lineif, at any instant, we have the data structure corresponding to the input we have seen up to that instant. For instance, consider the suffix tree of a stringx[1,n]. An algorithm building iton-lineis such that, when we have read the firstisymbols ofx[1,n], we have the suffix tree forx[1,i]. We present a new technique, which we refer to asimplicit updates, based on which we obtain: (a) an algorithm for theon-lineconstruction of the Lsuffix tree of ann×nmatrixA—this data structure is the two-dimensional analog of the suffix tree of a string; (b) simple algorithms implementing primitive operations forLZ1-typeon-line losslessimage compression m…
Parametric estimation of non-crossing quantile functions
2021
Quantile regression (QR) has gained popularity during the last decades, and is now considered a standard method by applied statisticians and practitioners in various fields. In this work, we applied QR to investigate climate change by analysing historical temperatures in the Arctic Circle. This approach proved very flexible and allowed to investigate the tails of the distribution, that correspond to extreme events. The presence of quantile crossing, however, prevented using the fitted model for prediction and extrapolation. In search of a possible solution, we first considered a different version of QR, in which the QR coefficients were described by parametric functions. This alleviated th…
Liquidity-adjusted value-at-risk optimization of a multi-asset portfolio using a vine copula approach
2019
Abstract This paper develops a novel approach to assess liquidity-adjusted Value-at-Risk (LVaR) optimization of multi-asset portfolios based on vine copulas and LVaR models. This framework is applied to stock markets of the G-7 countries, gold, commodities and Bitcoin. The results show that our approach is superior to the classical mean–variance Markowitz portfolio technique in terms of the optimal portfolio selection under a number of realistic operational and budget constraints. We find that both Bitcoin and gold improves the risk-return performance of the G-7 stock portfolio. However, Bitcoin (gold) performs better under a scenario of only long-positions (when short-selling is allowed).
A multi-local optimization algorithm
1998
The development of efficient algorithms that provide all the local minima of a function is crucial to solve certain subproblems in many optimization methods. A “multi-local” optimization procedure using inexact line searches is presented, and numerical experiments are also reported. An application of the method to a semi-infinite programming procedure is included.
The rank of random regular digraphs of constant degree
2018
Abstract Let d be a (large) integer. Given n ≥ 2 d , let A n be the adjacency matrix of a random directed d -regular graph on n vertices, with the uniform distribution. We show that the rank of A n is at least n − 1 with probability going to one as n grows to infinity. The proof combines the well known method of simple switchings and a recent result of the authors on delocalization of eigenvectors of A n .
The Serial Property and Restricted Balanced Contributions in discrete cost sharing problems
2006
We show that the Serial Poperty and Restricted Balanced Contributions characterize the subsidy-free serial cost sharing method (Moulin (1995)) in discrete cost allocation problems.
Hores: A timetabling system for Spanish secondary schools
1995
Constructing a timetable is a difficult problem faced by every school every year. A feasible solution has to satisfy many different requirements and constraints. A good solution has to provide compact timetables for classes and teachers. In order to help the schools, we have developed HORES, a robust and flexible timetabling system suited to the needs of Spanish secondary schools. HORES runs on a PC and is fast and user-friendly. It may handle virtually every condition required by the schools and obtains good quality solutions in very short computing times. It also allows the user to modify interactively the solutions. HORES is now being used by schools with satisfactory results.