خوارزميات التدرج المترافق للأمثلية غير المقيدة

الموضوعات

علوم بحتة وطبيعية

روابط وملفات

التعريف والنوع

رقم الوثيقة

fc9c417d-4de7-4163-b4f3-11c8e41627ad

رقم العقد

0

نوع الوسائط

Crawler

نوع المحتوى

الرسائل العلمية

صيغة المصدر

رسائل ماجيستير

نوع الملف

pdf text

أسماء الملفات

1959332_1.pdf

بيانات النشر

ترجمة العنوان

Conjugate Gradient Algorithms for Unconstrained Optimization

ألقاب المؤلفين

[{"name_ar":"زياد قنايه ","title_ar":"اشراف","title_en":"Supervision"},{"name_ar":"علي جمعه العيسى","title_ar":"اعداد","title_en":"Preparation"}]

اللغة

Arabic

المصدر والدورية

اسم المصدر

خوارزميات التدرج المترافق للأمثلية غير المقيدة

المحتوى والصفحات

عدد الصفحات

0

ترجمة الملخص

Optimization is considered one of the most widespread science as it is indulged in some fields such as economics, management, planning, supply, automatic control, engineering and guidance in addition to decision making processes. In this paper, it has been developed and enhanced a conjugate gradient algorithm by using a novel technique of choosing conjugate coefficient B in dependence on RMIL [1] and RMIL+[2]. The numerical solution of the following unconstrained optimization problem: That formula is mathematically equivalent to the following :subject to where is non-linear function This Thesis is written in Arabic language with English-Arabic abstract. It is consisted of three chapters with an index of contents and a list of scientific terminologies. The first chapter :It includes some important definitions and main concepts related to the unconstrained optimizations. It also provides some common algorithms to solve the unconstrained optimization problems, including the steepest descent, Newton algorithm, quasi-Newton algorithm, Davidon-Fletcher-Powell algorithm, Broyden-Fletcher-Goldfarb-Shanno algorithm, Golden search algorithm, Hook-Jeevessearch algorithm, Powell quadratic interpolation algorithm and cubic search algorithm. The second chapter: This chapter focuses on studying some traditional conjugate gradient algorithms such as Fletcher-Revers algorithm, conjugate gradient algorithm similar to PRP algorithm, RMIL algorithm, RMIL+ algorithm and ABarzilai- Borwein algorithm. The third chapter: 71This chapter presents a new algorithm to solve the unconstrained optimization problems. This suggested algorithm depends on developing conjugate coefficient by a hybridization between RMIL and RMIL+. However, the convergence of the suggested algorithm has been studied by proving the sufficient descent condition. This sufficient descent condition makes the numerical solution converged to the optimized one at each frequency, and it also helps in studying the comprehensive convergence of the suggested algorithm. The suggested algorithm has been examined by being applied on a group of measuring problems found in new studies. Then, a comparisons have been between our algorithm’s results and others’ ones in way the numerical results have showed the efficiency and the applicability effectiveness of the suggested algorithm. Noteworthy to mention that the numerical results have been got by such programs designed by using Matlab language. Some results of the third chapter have been approved and published by Tishreen University magazine for studies and scientific research.

إشراف وإعداد

الإشراف

زياد قنايه

الإعداد

علي جمعه العيسى

الاقتباسات الببليوغرافية