Results 11  20
of
33
ASAGA: An Adaptive SurrogateAssisted Genetic Algorithm
"... Genetic algorithms (GAs) used in complex optimization domains usually need to perform a large number of fitness function evaluations in order to get nearoptimal solutions. In real world application domains such as the engineering design problems, such evaluations might be extremely expensive comput ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
(Show Context)
Genetic algorithms (GAs) used in complex optimization domains usually need to perform a large number of fitness function evaluations in order to get nearoptimal solutions. In real world application domains such as the engineering design problems, such evaluations might be extremely expensive computationally. It is therefore common to estimate or approximate the fitness using certain methods. A popular method is to construct a so called surrogate or metamodel to approximate the original fitness function, which can simulate the behavior of the original fitness function but can be evaluated much faster. It is usually difficult to determine which approximate model should be used and/or what the frequency of usage should be. The answer also varies depending on the individual problem. To solve this problem, an adaptive fitness approximation GA (ASAGA) is presented. ASAGA adaptively chooses the appropriate model type; adaptively adjusts the model complexity and the frequency of model usage according to time spent and model accuracy. ASAGA also introduces a stochastic penalty function method to handle constraints. Experiments show that ASAGA outperforms nonadaptive surrogateassisted GAs with statistical significance.
Handling Constrained Optimization Problems and Using Constructive Induction to Improve Representation Spaces in Learnable Evolution Model
, 2007
"... iii ..."
(Show Context)
Constrained MultiObjective GA Optimization Using Reduced Models
"... In this paper we propose a novel approach for solving constrained multiobjective optimization problems using a steady state GA and reduced models. Our method called Objective Exchange Genetic Algorithm for Design optimization (OEGADO) is intended for solving realworld application problems that hav ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
In this paper we propose a novel approach for solving constrained multiobjective optimization problems using a steady state GA and reduced models. Our method called Objective Exchange Genetic Algorithm for Design optimization (OEGADO) is intended for solving realworld application problems that have many consmints and very small feasible regions. OEGADO runs several GAs concurrently with each GA optimizing one objective and exchanging information about its objective with others. Empirical results in benchmark and engineering design domains are presented. A comparison between OEGADO and NonDominated Sorting Genetic AlgorithmII (NSGAII) shows that OEGADO performed better than NSGAII for difficult problems, and found Paretooptimal solutions in fewer objective evaluations. The results suggest that our method may be better for solving realworld application problems wherein the objective computation time is large.
Validating Learnable Evolution Model on Selected Optimization and Design Problems
 George Mason University
, 2003
"... The recently introduced Learnable Evolution Model (LEM) represents a form of nonDarwinian evolutionary computation that is guided by a learning system. Specifically, LEM "genetically engineers" new populations via hypothesis formation and instantiation. Initial experiments with a prelimin ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
(Show Context)
The recently introduced Learnable Evolution Model (LEM) represents a form of nonDarwinian evolutionary computation that is guided by a learning system. Specifically, LEM "genetically engineers" new populations via hypothesis formation and instantiation. Initial experiments with a preliminary implementation of LEM were highly encouraging, but tentative. This paper presents results from a new study in which LEM was systematically tested on a range of optimization problems and a complex real world design task. The study involved LEM2, a new implementation oriented toward function optimization, and ISHED, an implementation oriented toward engineering design. In all cases of function optimization, LEM2 strongly outperformed tested evolutionary algorithms in terms of the evolution length, measured by the number of fitness function evaluations needed to reach the solution. This evolutionary speedup also translated to an execution speedup whenever the fitness evaluation time was above a small threshold (a fraction of a second). The most important result of the study was that the advantage of LEM2 over the tested Darwinianstyle evolutionary methods grew rapidly with the growth of the complexity of the optimized function. Experiments with ISHED on problems of optimizing heat exchangers (evaporators) produced designs that matched or exceeded designs produced by human experts. The obtained very strong results from the LEM application to two diverse domains suggest that it may be useful also in other application domains, especially, those in which the fitness function evaluation is timeconsuming or complex.
Abstract Determining the Optimal Cross Section of Beams
"... Constrained shape discovery and optimisation are difficult engineering problems. Shape discovery deals with evolving randomly generated solutions into useful intermediate solutions. These intermediate solutions are then optimised to suite their environment. In this paper a Genetic Algorithm (GA) is ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
Constrained shape discovery and optimisation are difficult engineering problems. Shape discovery deals with evolving randomly generated solutions into useful intermediate solutions. These intermediate solutions are then optimised to suite their environment. In this paper a Genetic Algorithm (GA) is applied to the problem of finding the optimum cross section of a beam, subject to various loading conditions. Previous work using GAs for this problem has relied heavily on heuristics and domain knowledge that operates directly on the genotype to guide the search. It is sound engineering practice to utilise all available information to reduce design and evaluation times, and encourage the formulation of useful solutions. However, domain design knowledge is not always available. This research attempts to explore the efficiency and effectiveness of a GA, when applied to a difficult design task, without being unnecessarily constrained by preconceptions of how to solve the task. Heavy guidance of a GA potentially stifles innovation, can only be applied to situations where the correct answer is known and limits the generic abilities of the search system. This research demonstrates the ability of the GA to evolve good, near optimal solutions without direct guidance. Performing an unbiased search, using only the evolutionary process to search for good solutions, allows a GA to be applied with a high degree of confidence to situations where a priori knowledge of the optimum solution is unavailable. Advanced 2dimensional genetic operators, in conjunction with a suitably designed fitness function, allow a productive evolutionary search. The initial test case is the evolution of an optimal Ibeam crosssection, subject to several load cases, starting with an initial random population. It is shown that the methods developed lead to consistently good solutions, despite the complexity of the process.
Nomenclature
, 2010
"... an = coefficient for parabolic rear shape a1, a2, b1, b2, c1, c2, d1, d2 = coefficients of cubic splines that parameterize the center portion of the envelope CD = drag coefficient d = envelope diameter l = envelope length n = load per unit length along meridians, i.e., in warp direction n ’ = load ..."
Abstract
 Add to MetaCart
(Show Context)
an = coefficient for parabolic rear shape a1, a2, b1, b2, c1, c2, d1, d2 = coefficients of cubic splines that parameterize the center portion of the envelope CD = drag coefficient d = envelope diameter l = envelope length n = load per unit length along meridians, i.e., in warp direction n ’ = load per unit length along latitude circles, i.e., in weft direction pR = internal overpressure in the aerostat envelope R = radius of curvature of spherical front portion of envelope R1, R2 = radii of curvature and transverse curvature of an inflated structure XD = design vector for shape optimization I.
Utility Guided Search of Stochastically Generated Trees
"... GAs have been found to be useful in handling many numerical optimization problems. Because of the variability in results inherent in the stochastic nature of GAs, it is common to run a GA several times and take the best of the results. However, it is possible to save a GA’s population at some interm ..."
Abstract
 Add to MetaCart
(Show Context)
GAs have been found to be useful in handling many numerical optimization problems. Because of the variability in results inherent in the stochastic nature of GAs, it is common to run a GA several times and take the best of the results. However, it is possible to save a GA’s population at some intermediate states and restart from one of these populations instead of from the very beginning. By doing so we generate a tree of populations, where a child population is generated from its parent by running some number of GA iterations. We describe two methods for searching such a tree of populations, one based on Highest Utility First Search (HUFS) and one that proceeds level by level with no backtracking, and give the results of testing them on a realworld optimization task involving conceptual design of supersonic transport aircraft. They both do much better than repeatedly running the GA from the beginning, with HUFS achieving equivalent results in less than half the GA iterations in some situations.
A New Approach to Optimizing Complex Engineering Systems and its Application to Designing Heat Exchangers
"... A new method for optimizing complex engineering designs is presented that is based on the Learnable Evolution Model (LEM), a recently developed form of nonDarwinian evolutionary computation. Unlike conventional Darwiniantype methods that execute an unguided evolutionary process, the proposed metho ..."
Abstract
 Add to MetaCart
(Show Context)
A new method for optimizing complex engineering designs is presented that is based on the Learnable Evolution Model (LEM), a recently developed form of nonDarwinian evolutionary computation. Unlike conventional Darwiniantype methods that execute an unguided evolutionary process, the proposed method, called LEMd, guides the evolutionary design process using a combination of two methods, one involving computational intelligence and the other involving encoded expert knowledge. Specifically, LEMd integrates two modes of operation, Learning Mode and Probing Mode. Learning Mode applies a machine learning program to create new designs through hypothesis generation and instantiation, while Probing Mode creates them by applying expertsuggested design modification operators tailored to the specific design problem. The LEMd method has been used to implement two initial systems, ISHED1 and ISCOD1, specialized for the optimization of evaporators and condensers in cooling systems, respectively. The designs produced by these systems matched or exceeded in performance the best designs developed by human experts. These promising results and the generality of the presented method suggest that LEMd offers a powerful new tool for optimizing complex engineering systems.