東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Developing New High-Order Sequential...

Yao, Lingqing,

Linked to FindBook

Google Book

Amazon

博客來

Developing New High-Order Sequential Simulation Methods Based on Learning-Oriented Kernels /

Record Type:	Electronic resources : Monograph/item
Title/Author:	Developing New High-Order Sequential Simulation Methods Based on Learning-Oriented Kernels // Lingqing Yao.
Author:	Yao, Lingqing,
Description:	1 electronic resource (180 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 83-08, Section: B.
Contained By:	Dissertations Abstracts International83-08B.
Subject:	Gold. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28990241
ISBN:	9798759977681

Developing New High-Order Sequential Simulation Methods Based on Learning-Oriented Kernels /
Yao, Lingqing,

Developing New High-Order Sequential Simulation Methods Based on Learning-Oriented Kernels /Lingqing Yao. - 1 electronic resource (180 pages)

Source: Dissertations Abstracts International, Volume: 83-08, Section: B.

Uncertainty quantification plays a vital role in managing the technical risk of the sustainable exploitation of natural resources. Random field models are utilized to model the natural attributes of interest, within which the attributes at different locations are represented as random variables comprising a joint probability distribution. Spatial statistics, varied with different random field models, mathematically describe the spatial patterns. Stochastic simulation methods generate multiple realizations based on certain random field models to represent the possible outcomes of natural attributes under consideration. They aim to reproduce spatial statistics of the perceived data, thus providing useful tools to quantify the spatial uncertainty of the target attributes. In the context of mining applications, reproducing spatial patterns from the sample data has a significant impact on managing the risks of mine planning decisions. Specifically, the net present value (NPV) regarding a certain mine planning schedule of a mineral deposit depends on the revenue generated by the extraction sequences of the underground materials, as the cash flows are discounted by the mining periods. The extraction sequences of the materials, in turn, are driven by the spatial distributions of metal grades, especially the spatial continuity of enriched metal elements. High-order stochastic simulation methods make no assumption of any specific probability distribution on the random field models, avoiding the limitation of traditional Gaussian random field models. In addition, the methods account for the high-order spatial statistics that characterize the statistical interactions among random attributes at multiple locations and thus have the advantage of reproducing complex spatial patterns. Therefore, this thesis develops new high-order stochastic simulation methods based on a proposed framework of statistical learning and learning-oriented kernels, aiming to advance the theoretical aspects of the stochastic simulation methods, as well as the practical aspects of mining decisions under uncertainty. The general paradigm of sequential simulation is adopted in this thesis to generate realizations from the random field models, which decomposes the joint probability distributions into a sequence of conditional probability distributions.

English

ISBN: 9798759977681Subjects--Topical Terms:

659820
Gold.

Developing New High-Order Sequential Simulation Methods Based on Learning-Oriented Kernels /
LDR:06542nmm a22004333i 4500 001 2437620
005 20251124054947.5
006 m o d
007 cr|nu||||||||
008 261011s2020 miu||||||m |||||||eng d
020 $a 9798759977681
035 $a (MiAaPQD)AAI28990241
035 $a (MiAaPQD)EPMontreal5342
035 $a AAI28990241
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Yao, Lingqing, $e author. $3 3828671
245 1 0 $a Developing New High-Order Sequential Simulation Methods Based on Learning-Oriented Kernels / $c Lingqing Yao.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2020
300 $a 1 electronic resource (180 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 83-08, Section: B.
500 $a Advisors: Gamache, Michel; Dimitrakopoulos, Roussos.
502 $b Ph.D. $c Ecole Polytechnique, Montreal (Canada) $d 2020.
520 $a Uncertainty quantification plays a vital role in managing the technical risk of the sustainable exploitation of natural resources. Random field models are utilized to model the natural attributes of interest, within which the attributes at different locations are represented as random variables comprising a joint probability distribution. Spatial statistics, varied with different random field models, mathematically describe the spatial patterns. Stochastic simulation methods generate multiple realizations based on certain random field models to represent the possible outcomes of natural attributes under consideration. They aim to reproduce spatial statistics of the perceived data, thus providing useful tools to quantify the spatial uncertainty of the target attributes. In the context of mining applications, reproducing spatial patterns from the sample data has a significant impact on managing the risks of mine planning decisions. Specifically, the net present value (NPV) regarding a certain mine planning schedule of a mineral deposit depends on the revenue generated by the extraction sequences of the underground materials, as the cash flows are discounted by the mining periods. The extraction sequences of the materials, in turn, are driven by the spatial distributions of metal grades, especially the spatial continuity of enriched metal elements. High-order stochastic simulation methods make no assumption of any specific probability distribution on the random field models, avoiding the limitation of traditional Gaussian random field models. In addition, the methods account for the high-order spatial statistics that characterize the statistical interactions among random attributes at multiple locations and thus have the advantage of reproducing complex spatial patterns. Therefore, this thesis develops new high-order stochastic simulation methods based on a proposed framework of statistical learning and learning-oriented kernels, aiming to advance the theoretical aspects of the stochastic simulation methods, as well as the practical aspects of mining decisions under uncertainty. The general paradigm of sequential simulation is adopted in this thesis to generate realizations from the random field models, which decomposes the joint probability distributions into a sequence of conditional probability distributions.
520 $a La quantification de l'incertitude joue un role essentiel dans la gestion du risque technique de l'exploitation durable des ressources naturelles. Les modeles de champs aleatoires sont utilises pour modeliser les attributs naturels d'interet, parmi lesquels les attributs a differents endroits sont representes comme des variables aleatoires comprenant une distribution de probabilite conjointe. Les statistiques spatiales, qui varient selon differents modeles de champs aleatoires, decrivent mathematiquement les structures spatiales. Les methodes de simulation stochastique generent des realisations multiples basees sur certains modeles de champs aleatoires afin de representer les resultats possibles des attributs naturels consideres. Elles visent a reproduire les statistiques spatiales des donnees percues, fournissant ainsi des outils utiles pour quantifier l'incertitude spatiale des attributs cibles. Dans le contexte des applications minieres, la reproduction de structures spatiales, a partir des donneesechantillons, a un impact significatif sur la gestion des risques lies aux decisions de planification miniere. Plus precisement, la valeur actualisee nette (VAN) d'un gisement mineral, compte tenu d'un calendrier de planification miniere donne, depend des revenus generes par les sequences d'extraction des materiaux souterrains, les flux de tresorerie etant actualises en fonction des periodes d'exploitation. Les sequences d'extraction des materiaux, a leur tour, sont determinees par la distribution spatiale des teneurs en metaux, en particulier la continuite spatiale des elements metalliques enrichis. Les methodes de simulation stochastique d'ordre eleve ne presupposent aucune distribution de probabilite specifique sur les modeles de champs aleatoires, evitant ainsi les limites des modeles de champs aleatoires gaussiens traditionnels. De plus, ces methodes tiennent compte des statistiques spatiales d'ordre eleve qui caracterisent les interactions statistiques entre les attributs aleatoires en de multiples endroits et elles ont donc l'avantage de reproduire des structures spatiales complexes. Par consequent, cette these developpe de nouvelles methodes de simulation stochastique d'ordre eleve basees sur un cadre propose d'apprentissage statistique et de noyaux orientes sur l'apprentissage, visant a faire progresser les aspects theoriques des methodes de simulation stochastique ainsi que les aspects pratiques des decisions minieres sous incertitude. Le paradigme general de la simulation sequentielle est adopte dans cette these afin de generer des realisations a partir de modeles de champs aleatoires, ce qui decompose les distributions de probabilites conjointes en une sequence de distributions de probabilites conditionnelles.
546 $a English
590 $a School code: 1105
650 4 $a Gold. $3 659820
650 4 $a Mining. $3 3544442
650 4 $a Industrial engineering. $3 526216
650 4 $a Statistics. $3 517247
650 4 $a Natural resource management. $3 589570
650 4 $a Mineral resources. $3 3707173
650 4 $a Sustainability. $3 1029978
690 $a 0551
690 $a 0640
690 $a 0546
690 $a 0528
690 $a 0463
710 2 $a Ecole Polytechnique, Montreal (Canada). $e degree granting institution. $3 3821806
720 1 $a Gamache, Michel $e degree supervisor.
720 1 $a Dimitrakopoulos, Roussos $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 83-08B.
790 $a 1105
791 $a Ph.D.
792 $a 2020
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28990241