Naslov
Two-point Mid-range Approximation Enhanced Recursive Quadratic Programming Method

Autori
Magazinović, Gojko

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Short Papers of the Fifth World Congress of Structural and Multidisciplinary Optimization / Cinquini, Carlo ; Rovati, Marco ; Venini, Paolo ; Nascimbene, Roberto - Milano : Italian Polytechnic Press, 2003, 189-190

Skup
Fifth World Congress of Structural and Multidisciplinary Optimization

Mjesto i datum
Lido di Jesolo, Italija, 19.05.2003. - 23.05.2003

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
design optimization; function approximation; recursive quadratic programming

Sažetak
This research presents an attempt to combine good convergence properties of recursive quadratic programming methods with the benefits of mid-range approximations, initially developed in the field of structural optimization. In this paper, an optimization method based on Arora and coworkers PLBA (Pshenichny-Lim-Belegundu-Arora) algorithm is proposed in which, during the line search phase, cost and constraint functions are substituted by their two-point approximations using the Generalized Convex Approximation formulae of Chickermane and Gea. The results showed that the proposed optimization method preserves the reliability and accuracy of the recursive quadratic programming method while it might simultaneously reduce the computational effort on some problems. Therefore, the proposed optimization method may be taken as potentially suitable for general design optimization purposes.

Izvorni jezik
Engleski

Znanstvena područja
Strojarstvo

POVEZANOST RADA