STUDY OF HYBRID STRATEGIES FOR MULTI-OBJECTIVE OPTIMIZATION USING GRADIENT BASED METHODS AND EVOLUTIONARY ALGORITHMS

Search for this publication on Google Scholar
Paez Bautista, D. (2013). STUDY OF HYBRID STRATEGIES FOR MULTI-OBJECTIVE OPTIMIZATION USING GRADIENT BASED METHODS AND EVOLUTIONARY ALGORITHMS. Unc Charlotte Electronic Theses And Dissertations.
Download PDF

Analytics

62 views ◎
29 downloads ⇓


Abstract

Most of the optimization problems encountered in engineering have conflicting objectives. In order to solve these problems, genetic algorithms (GAs) and gradient-based methods are widely used. GAs are relatively easy to implement, because these algorithms only require first-order information of the objectives and constraints. On the other hand, GAs do not have a standard termination condition and therefore they may not converge to the exact solutions. Gradient-based methods, on the other hand, are based on first- and higher-order information of the objectives and constraints. These algorithms converge faster to the exact solutions in solving single-objective optimization problems, but are inefficient for multi-objective optimization problems (MOOPs) and unable to solve those with non-convex objective spaces.The work in this dissertation focuses on developing a hybrid strategy for solving MOOPs based on feasible sequential quadratic programming (FSQP) and nondominated sorting genetic algorithm II (NSGA-II). The hybrid algorithms developed in this dissertation are tested using benchmark problems and evaluated based on solution distribution, solution accuracy, and execution time. Based on these performance factors, the best hybrid strategy is determined and found to be generally efficient with good solution distributions in most of the cases studied. The best hybrid algorithm is applied to the design of a crushing tube and is shown to have relatively well-distributed solutions and good efficiency compared to solutions obtained by NSGA-II and FSQP alone.

Details
Author

Paez Bautista, Diego

Title

STUDY OF HYBRID STRATEGIES FOR MULTI-OBJECTIVE OPTIMIZATION USING GRADIENT BASED METHODS AND EVOLUTIONARY ALGORITHMS

Physical Description

1 online resource (192 pages) : PDF

Date

2013

Degree Granting Institution

University of North Carolina at Charlotte

Abstract

Most of the optimization problems encountered in engineering have conflicting objectives. In order to solve these problems, genetic algorithms (GAs) and gradient-based methods are widely used. GAs are relatively easy to implement, because these algorithms only require first-order information of the objectives and constraints. On the other hand, GAs do not have a standard termination condition and therefore they may not converge to the exact solutions. Gradient-based methods, on the other hand, are based on first- and higher-order information of the objectives and constraints. These algorithms converge faster to the exact solutions in solving single-objective optimization problems, but are inefficient for multi-objective optimization problems (MOOPs) and unable to solve those with non-convex objective spaces.The work in this dissertation focuses on developing a hybrid strategy for solving MOOPs based on feasible sequential quadratic programming (FSQP) and nondominated sorting genetic algorithm II (NSGA-II). The hybrid algorithms developed in this dissertation are tested using benchmark problems and evaluated based on solution distribution, solution accuracy, and execution time. Based on these performance factors, the best hybrid strategy is determined and found to be generally efficient with good solution distributions in most of the cases studied. The best hybrid algorithm is applied to the design of a crushing tube and is shown to have relatively well-distributed solutions and good efficiency compared to solutions obtained by NSGA-II and FSQP alone.

Genre

doctoral dissertations

Subjects--Topics

Mechanical engineering
Mathematics

Degree

Ph.D.

Keywords

Engineering Optimization
Genetic Algorithms
Gradient-Based Optimization
Hybrid Algorithms
Hybrid Optimization
Non-Convex Optimization

Subject Area

Mechanical Engineering

Advisor(s)

Fang, Howie

Committee Members

Keanini, Russell
Srivastava, Nilabh
Young, David
Weggel, David

Degree Note

Thesis (Ph.D.)--University of North Carolina at Charlotte, 2013.

Rights Statement

This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s). For additional information, see http://rightsstatements.org/page/InC/1.0/.

Rights Holder Information

Copyright is held by the author unless otherwise indicated.

Identifier

PaezBautista_uncc_0694D_10495

Permalink

http://hdl.handle.net/20.500.13093/etd:180