FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3119736473> ?p ?o ?g. }

• W3119736473 abstract "Industrial problems often take the form of expensive laboratory experiments or computationally-heavy simulations. These expensive experiments or simulations provide valuable information about the underlying latent physical process being modeled. One is often faced with the problem of optimizing the quantity of interest (QoI) being represented by the information source. Querying the information source tens of thousands of times, generally required by state-of-the-art optimization algorithms, is practically infeasible. The problem becomes cumbersome when the number of objectives being optimized is greater than one. In such multi-objective optimization (MOO) scenarios, expected improvement (EI)-based methods have demonstrated great efficiency in converging to the optimal solution while taking into consideration a fixed budget. The EI-based methods, used for expensive MOO problems, are broadly called EI-hypervolume (EIHV) methods. The EI-based algorithm called, Bayesian global optimization (BGO), builds cheap probabilistic models of the objectives, using Gaussian process regression(GPR), and selects the next input to query the information source from amongst a finite set of candidate inputs. The EIHV weighs the potential improvement associated with each of the candidate designs. This iterative process continues until one exhausts the budget. The evaluation of the EIHV becomes painfully expensive, scales exponentially, as the number of objectives increases. The finite set of possible inputs, used to evaluate the EIHV and select the next query, ceases to be a computationally feasible option in order for the overall optimization to proceed. In this work, we aim to investigate the performance of an EI-based algorithm, for optimizing the EIHV for MOO problems. At each iteration of the optimization problem, we optimize the EIHV using EI-based BGO, in order to be computationally efficient. We demonstrate the approach on two mathematical MOO problems and one challenging industrial problem with four objectives. We also compare the results in terms of convergence and savings in computational time with the finite-grid based approach." @default.
• W3119736473 created "2021-01-18" @default.
• W3119736473 creator A5025062741 @default.
• W3119736473 creator A5034885122 @default.
• W3119736473 creator A5052787170 @default.
• W3119736473 date "2021-01-04" @default.
• W3119736473 modified "2023-09-23" @default.
• W3119736473 title "Computationally Efficient Bayesian Optimization for Multi-objective Industrial Applications" @default.
• W3119736473 cites W1487465581 @default.
• W3119736473 cites W1510052597 @default.
• W3119736473 cites W1542433298 @default.
• W3119736473 cites W2027531431 @default.
• W3119736473 cites W2032558547 @default.
• W3119736473 cites W2063180182 @default.
• W3119736473 cites W2075106202 @default.
• W3119736473 cites W2098629435 @default.
• W3119736473 cites W2102365077 @default.
• W3119736473 cites W2126105956 @default.
• W3119736473 cites W2136918060 @default.
• W3119736473 cites W2144728442 @default.
• W3119736473 cites W2151238122 @default.
• W3119736473 cites W2156267707 @default.
• W3119736473 cites W2395865283 @default.
• W3119736473 cites W2753043552 @default.
• W3119736473 cites W2781579237 @default.
• W3119736473 cites W2781701821 @default.
• W3119736473 cites W4249517230 @default.
• W3119736473 cites W4252003911 @default.
• W3119736473 cites W85584663 @default.
• W3119736473 doi "https://doi.org/10.2514/6.2021-1482" @default.
• W3119736473 hasPublicationYear "2021" @default.
• W3119736473 type Work @default.
• W3119736473 sameAs 3119736473 @default.
• W3119736473 citedByCount "0" @default.
• W3119736473 crossrefType "proceedings-article" @default.
• W3119736473 hasAuthorship W3119736473A5025062741 @default.
• W3119736473 hasAuthorship W3119736473A5034885122 @default.
• W3119736473 hasAuthorship W3119736473A5052787170 @default.
• W3119736473 hasConcept C107673813 @default.
• W3119736473 hasConcept C111919701 @default.
• W3119736473 hasConcept C11413529 @default.
• W3119736473 hasConcept C119857082 @default.
• W3119736473 hasConcept C121332964 @default.
• W3119736473 hasConcept C124101348 @default.
• W3119736473 hasConcept C126255220 @default.
• W3119736473 hasConcept C137836250 @default.
• W3119736473 hasConcept C154945302 @default.
• W3119736473 hasConcept C163716315 @default.
• W3119736473 hasConcept C177264268 @default.
• W3119736473 hasConcept C199360897 @default.
• W3119736473 hasConcept C2778049539 @default.
• W3119736473 hasConcept C33923547 @default.
• W3119736473 hasConcept C41008148 @default.
• W3119736473 hasConcept C49937458 @default.
• W3119736473 hasConcept C61326573 @default.
• W3119736473 hasConcept C62520636 @default.
• W3119736473 hasConcept C81692654 @default.
• W3119736473 hasConcept C98045186 @default.
• W3119736473 hasConceptScore W3119736473C107673813 @default.
• W3119736473 hasConceptScore W3119736473C111919701 @default.
• W3119736473 hasConceptScore W3119736473C11413529 @default.
• W3119736473 hasConceptScore W3119736473C119857082 @default.
• W3119736473 hasConceptScore W3119736473C121332964 @default.
• W3119736473 hasConceptScore W3119736473C124101348 @default.
• W3119736473 hasConceptScore W3119736473C126255220 @default.
• W3119736473 hasConceptScore W3119736473C137836250 @default.
• W3119736473 hasConceptScore W3119736473C154945302 @default.
• W3119736473 hasConceptScore W3119736473C163716315 @default.
• W3119736473 hasConceptScore W3119736473C177264268 @default.
• W3119736473 hasConceptScore W3119736473C199360897 @default.
• W3119736473 hasConceptScore W3119736473C2778049539 @default.
• W3119736473 hasConceptScore W3119736473C33923547 @default.
• W3119736473 hasConceptScore W3119736473C41008148 @default.
• W3119736473 hasConceptScore W3119736473C49937458 @default.
• W3119736473 hasConceptScore W3119736473C61326573 @default.
• W3119736473 hasConceptScore W3119736473C62520636 @default.
• W3119736473 hasConceptScore W3119736473C81692654 @default.
• W3119736473 hasConceptScore W3119736473C98045186 @default.
• W3119736473 hasLocation W31197364731 @default.
• W3119736473 hasOpenAccess W3119736473 @default.
• W3119736473 hasPrimaryLocation W31197364731 @default.
• W3119736473 hasRelatedWork W2032602526 @default.
• W3119736473 hasRelatedWork W2122656899 @default.
• W3119736473 hasRelatedWork W2507981315 @default.
• W3119736473 hasRelatedWork W2950792054 @default.
• W3119736473 hasRelatedWork W3026715030 @default.
• W3119736473 hasRelatedWork W3175956409 @default.
• W3119736473 hasRelatedWork W4210726438 @default.
• W3119736473 hasRelatedWork W4286492427 @default.
• W3119736473 hasRelatedWork W4287776349 @default.
• W3119736473 hasRelatedWork W4292081304 @default.
• W3119736473 isParatext "false" @default.
• W3119736473 isRetracted "false" @default.
• W3119736473 magId "3119736473" @default.
• W3119736473 workType "article" @default.