Pengaturcaraan Tak Linear Berpemangkin

Abstract

Antara banyak kaedah bagi menyelesaikan masalah pengoptimuman berkekangan tak linear, kaedah yang amat berkesan dalam konteks pengujiaannya ke atas masalah-masalah piawai melibatkan sama ada lelaran linear kuadratik. Maka, kertas kerja ini mencadangkan satu kaedah yang bukan sahaja menggunakan lelaran linear dan juga kuadratik tetapi menunjukkan kaedah cerun conjugat pelbagai rumus bagi pengoptimuman bebas dan penyelesaian submasalah pengaturcaraan linear sebelum rutin Newton dipanggil untuk memenuhi persamaan-persamaan kekangan. Katakunci: pengaturcaraan tak linear; kaedah cerun conjugat; pengoptimuman; rutin Newton Among the various methods of solving nonlinear programming problems, the effective ones in the context of solving several standard problems involve either linearization or quadratic iterations. This paper thus suggests a method which does not only use linearization or make quadratict approximations but shows how the nonlinear programming process can be catalyzed by using formulae variable conjugate gradient stratigies for the unconstrained potimization and solving the linear programming subproblem before the Newton routine is called so as to satisfy the constrain equations. Keywords: nonlinear programming; variable conjugate gradient; optimization; Newton routine