Distributed Newton-Seeking for Resource Allocation

dc.contributor.authorSchultz, Matten
dc.contributor.departmentChemical Engineeringen
dc.contributor.supervisorGuay, Martin
dc.date.accessioned2022-08-11T18:33:47Z
dc.date.available2022-08-11T18:33:47Z
dc.degree.grantorQueen's University at Kingstonen
dc.description.abstractReal-time optimization (RTO) is a category of closed-loop process control techniques that is focused on the improvement of process performance. To develop effective strategies, models of the systems are usually required. To circumvent this issue, model-free techniques such as Extremum-seeking control (ESC) have been developed that can accomplish the RTO task in the absence of any knowledge of the process dynamics and of the objective function. Using this optimization methodology, this thesis develops an approach to solve resource allocation problems in a fully distributed fashion where the separable cost function is unknown with dynamics in discrete time. To accomplish this, we utilize a distributed extremum-seeking scheme that estimates the cost function’s gradient and Hessian to implement Newton step dynamics. It is shown that the extremum-seeking approach yields asymptotic convergence of the primal and dual problem to a neighbourhood of the optimal values. We propose two strategies to solve finite resource allocation problems. The first strategy tackles the globally convex cost function case. It requires a simultaneous solution of the primal problem and its associated dual problem. The second strategy utilizes a more general approach with an pseudo-Hessian for the solution of local saddle point problems. It allows a full separation of the primal and dual dynamics. Two simulation case studies are performed for each strategy. The first case is a small-scale 3-agent problem highlighting the differences between the strategies. The second is a 25-agent problem that mimics more realistic scenarios. Finally, a stability analysis is conducted to characterize the transient performance of the two strategies.en
dc.description.degreeM.A.Sc.en
dc.identifier.urihttp://hdl.handle.net/1974/30297
dc.language.isoengen
dc.relation.ispartofseriesCanadian thesesen
dc.rightsQueen's University's Thesis/Dissertation Non-Exclusive License for Deposit to QSpace and Library and Archives Canadaen
dc.rightsProQuest PhD and Master's Theses International Dissemination Agreementen
dc.rightsIntellectual Property Guidelines at Queen's Universityen
dc.rightsCopying and Preserving Your Thesisen
dc.rightsThis publication is made available by the authority of the copyright owner solely for the purpose of private study and research and may not be copied or reproduced except as permitted by the copyright laws without written authority from the copyright owner.en
dc.rightsAttribution-NonCommercial-ShareAlike 3.0 United States*
dc.rights.urihttp://creativecommons.org/licenses/by-nc-sa/3.0/us/*
dc.subjectDistributed Optimizationen
dc.subjectResource Allocationen
dc.subjectConsensusen
dc.subjectExtremum Seeking Controlen
dc.titleDistributed Newton-Seeking for Resource Allocationen
dc.typethesisen
Files
Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
Schultz_Matthew_T_202209_MASC.pdf
Size:
1.63 MB
Format:
Adobe Portable Document Format
Description:
Thesis document
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.67 KB
Format:
Item-specific license agreed upon to submission
Description: