Parallel Branch and Bound Method for Size Optimization Benchmarks

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21110%2F13%3A00205079" target="_blank" >RIV/68407700:21110/13:00205079 - isvavai.cz</a>

Výsledek na webu

<a href="http://www.ctresources.info/ccp/paper.html?id=7299" target="_blank" >http://www.ctresources.info/ccp/paper.html?id=7299</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.4203/ccp.101.20" target="_blank" >10.4203/ccp.101.20</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Parallel Branch and Bound Method for Size Optimization Benchmarks

Popis výsledku v původním jazyce

This contribution presents a new formulation of a classical Branch and Bound method to find global optima of size optimization benchmarks. In our previous procedure, the brunching was done at an assignment of areas to the trusses, i.e. we have been solving the problem in the integer manner. However, the classical method requires a model with convex objectives and constraints, which was not the case. To convert the non-convex problem to the convex one, a relaxation of some variables is necessary. Original design variables ? cross-section areas on truss-bars are converted to the binary design variables with the meaning whether a cross-section area is presented on the respective truss-bar or not. This problem is relaxed and gives us the lower bound. Moreover, a parallel version of the classical Branch and Bound algorithm was implemented, where the upgrades of the lower and upper bounds are broadcasted among individual processes. This algorithm was used to compute the global optimum for ou

Název v anglickém jazyce

Parallel Branch and Bound Method for Size Optimization Benchmarks

Popis výsledku anglicky

This contribution presents a new formulation of a classical Branch and Bound method to find global optima of size optimization benchmarks. In our previous procedure, the brunching was done at an assignment of areas to the trusses, i.e. we have been solving the problem in the integer manner. However, the classical method requires a model with convex objectives and constraints, which was not the case. To convert the non-convex problem to the convex one, a relaxation of some variables is necessary. Original design variables ? cross-section areas on truss-bars are converted to the binary design variables with the meaning whether a cross-section area is presented on the respective truss-bar or not. This problem is relaxed and gives us the lower bound. Moreover, a parallel version of the classical Branch and Bound algorithm was implemented, where the upgrades of the lower and upper bounds are broadcasted among individual processes. This algorithm was used to compute the global optimum for ou

Druh

D - Stať ve sborníku

CEP obor

JD - Využití počítačů, robotika a její aplikace

OECD FORD obor

—

Projekt

<a href="/cs/project/GAP105%2F12%2F1146" target="_blank" >GAP105/12/1146: Metody paralelizace inženýrských úloh využívající cenově dostupné technologie</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2013

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Proceedings of the Third International Conference on Parallel, Distributed, Grid and Cloud Computing for Engineering

ISBN

978-1-905088-56-0

ISSN

1759-3433

e-ISSN

—

Počet stran výsledku

18

Strana od-do

—

Název nakladatele

Civil-Comp Press Ltd

Místo vydání

Stirling

Místo konání akce

Pécs

Datum konání akce

25. 3. 2013

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—