Data Structures and Algorithms

1503 Submissions

[3] viXra:1503.0220 [pdf] submitted on 2015-03-28 06:17:24

Geometry Inspired Algorithms for Linear Programming

Authors: Dhananjay P. Mehendale
Comments: 17 pages.

In this paper we discuss some novel algorithms for linear programming inspired by geometrical considerations and use simple mathematics related to finding intersections of lines and planes. All these algorithms have a common aim: they all try to approach closer and closer to “centroid” or some “centrally located interior point” for speeding up the process of reaching an optimal solution! Imagine the “line” parallel to vector C, where CTx denotes the objective function to be optimized, and further suppose that this “line” is also passing through the “point” representing optimal solution. The new algorithms that we propose in this paper essentially try to reach at some feasible interior point which is in the close vicinity of this “line”, in successive steps. When one will be able to arrive finally at a point belonging to small neighborhood of some point on this “line” then by moving from this point parallel to vector C one can reach to the point belonging to the sufficiently small neighborhood of the “point” representing optimal solution.
Category: Data Structures and Algorithms

[2] viXra:1503.0218 [pdf] submitted on 2015-03-27 19:59:36

Fusion D'images Par la Théorie de Dezert-Smarandache (DSmT) en Vue D'applications en Télédétection (Thèse Doctorale)

Authors: Azeddine Elhassouny
Comments: 120 Pages.

Thèse dirigée par Pr. Driss Mammass, préparée au Laboratoire Image et Reconnaissance de Formes-Systèmes Intelligents et Communicants IRF-SIC, soutenue le 22 juin 2013, Agadir, Maroc. L'objectif de cette thèse est de fournir à la télédétection des outils automatiques de la classification et de la détection des changements d'occupation du sol utiles à plusieurs fins, dans ce cadre, nous avons développé deux méthodes générales de fusion utilisées pour la classification des images et la détection des changements utilisant conjointement l'information spatiale obtenue par la classification supervisée ICM et la théorie de Dezert-Smarandache (DSmT) avec des nouvelles règles de décision pour surmonter les limites inhérentes des règles de décision existantes dans la littérature. L'ensemble des programmes de cette thèse ont été implémentés avec MATLAB et les prétraitements et visualisation des résultats ont été réalisés sous ENVI 4.0, ceci a permis d'effectuer une validation des résultats avec précision et dans des cas concrets. Les deux approches sont évaluées sur des images LANDSAT ETM+ et FORMOSAT-2 et les résultats sont prometteurs. The main objective of this thesis is to provide automatic remote sensing tools of classification and of change detection of land cover for many purposes, in this context, we have developed two general methods used for classification fusion images and change detection using joint spatial information obtained by supervised classification ICM and Dezert-Smarandache theory (DSmT) with new decision rules to overcome the limitations of decision rules existing in the literature. All programs of this thesis have been implemented in MATLAB and C language and preprocessing and visualization of results were achieved in ENVI 4.0, this has allowed for a validation of the results accurately and in concrete cases. Both approaches are evaluated on LANDSAT ETM + and FORMOSAT-2 and the results are promising.
Category: Data Structures and Algorithms

[1] viXra:1503.0018 [pdf] replaced on 2015-03-03 12:52:40

Advances and Applications of DSmT for Information Fusion. Collected Works. Volume 4

Authors: editors Florentin Smarandache, Jean Dezert
Comments: 504 Pages.

The fourth volume on Advances and Applications of Dezert-Smarandache Theory (DSmT) for information fusion collects theoretical and applied contributions of researchers working in different fields of applications and in mathematics. The contributions (see List of Articles published in this book, at the end of the volume) have been published or presented after disseminating the third volume (2009, http://fs.gallup.unm.edu/DSmT-book3.pdf) in international conferences, seminars, workshops and journals. First Part of this book presents the theoretical advancement of DSmT, dealing with Belief functions, conditioning and deconditioning, Analytic Hierarchy Process, Decision Making, Multi-Criteria, evidence theory, combination rule, evidence distance, conflicting belief, sources of evidences with different importance and reliabilities, importance of sources, pignistic probability transformation, Qualitative reasoning under uncertainty, Imprecise belief structures, 2-Tuple linguistic label, Electre Tri Method, hierarchical proportional redistribution, basic belief assignment, subjective probability measure, Smarandache codification, neutrosophic logic, Evidence theory, outranking methods, Dempster-Shafer Theory, Bayes fusion rule, frequentist probability, mean square error, controlling factor, optimal assignment solution, data association, Transferable Belief Model, and others. More applications of DSmT have emerged in the past years since the apparition of the third book of DSmT 2009. Subsequently, the second part of this volume is about applications of DSmT in correlation with Electronic Support Measures, belief function, sensor networks, Ground Moving Target and Multiple target tracking, Vehicle-Born Improvised Explosive Device, Belief Interacting Multiple Model filter, seismic and acoustic sensor, Support Vector Machines, Alarm classification, ability of human visual system, Uncertainty Representation and Reasoning Evaluation Framework, Threat Assessment, Handwritten Signature Verification, Automatic Aircraft Recognition, Dynamic Data-Driven Application System, adjustment of secure communication trust analysis, and so on. Finally, the third part presents a List of References related with DSmT published or presented along the years since its inception in 2004, chronologically ordered.
Category: Data Structures and Algorithms