The Effect of Multi-Parent Recombination on Evolution Strategies for Noisy Objective Functions

2011 ◽  
pp. 264-280
Author(s):  
Yoshiyuki Matsumura ◽  
Kazuhiro Ohkura ◽  
Kanji Ueda
Keyword(s):  
Computer Simulations ◽  
Evolution Strategies ◽  
Test Functions ◽  
Objective Functions ◽  
The Many ◽  
Object Parameters ◽  
Strategy Parameters

In this chapter we apply (m / m, l)-ES to noisy test functions, in order to investigate the effect of multi-parent versions of both intermediate recombination and discrete recombination. Among the many formulations of ES, we test three in particular; Classical-ES (CES), i.e., Schwefel’s original ES (Schwefel, 1995, Bäck, 1996); Fast-ES (FES), i.e., Yao and Liu’s extended ES (Yao & Liu, 1997); and Robust-ES (RES), i.e., our extended ES (Ohkura, 2001). Computer simulations are used to compare the performance of multi-parent versions of intermediate recombination and discrete recombination in CES, FES and RES. We saw that the performance of the (m / m, l)-ES algorithms depended on the particular objective functions. However, the FES and RES algorithms were seen to be improved by multi-parent versions of discrete recombination applied to both object parameters and strategy parameters.

Convergence in Evolutionary Programs with Self-Adaptation

2001 ◽  
Vol 9 (2) ◽  
pp. 147-157 ◽  
Author(s):  
Garrison W. Greenwood ◽  
Qiji J. Zhu
Keyword(s):  
Optimization Problems ◽  
Previous Analysis ◽  
Evolution Strategies ◽  
Convergence Properties ◽  
Formal Proofs ◽  
Self Adaptation ◽  
Long Term Behavior ◽  
Adaptation Method ◽  
Strategy Parameters

Evolutionary programs are capable of finding good solutions to difficult optimization problems. Previous analysis of their convergence properties has normally assumed the strategy parameters are kept constant, although in practice these parameters are dynamically altered. In this paper, we propose a modified version of the 1/5-success rule for self-adaptation in evolution strategies (ES). Formal proofs of the long-term behavior produced by our self-adaptation method are included. Both elitist and non-elitist ES variants are analyzed. Preliminary tests indicate an ES with our modified self-adaptation method compares favorably to both a non-adapted ES and a 1/5-success rule adapted ES.

scholarly journals Using particle swarm optimization to solve test functions problems

2021 ◽  
Vol 10 (6) ◽  
pp. 3422-3431
Author(s):  
Issa Ahmed Abed ◽  
May Mohammed Ali ◽  
Afrah Abood Abdul Kadhim
Keyword(s):  
Genetic Algorithm ◽  
Particle Swarm Optimization ◽  
Standard Deviation ◽  
Particle Swarm ◽  
Pso Algorithm ◽  
Speed Of Convergence ◽  
Test Functions ◽  
Objective Functions ◽  
Swarm Optimization ◽  
Engineering Problems

In this paper the benchmarking functions are used to evaluate and check the particle swarm optimization (PSO) algorithm. However, the functions utilized have two dimension but they selected with different difficulty and with different models. In order to prove capability of PSO, it is compared with genetic algorithm (GA). Hence, the two algorithms are compared in terms of objective functions and the standard deviation. Different runs have been taken to get convincing results and the parameters are chosen properly where the Matlab software is used. Where the suggested algorithm can solve different engineering problems with different dimension and outperform the others in term of accuracy and speed of convergence.

Fractal Analysis of the Optimal Objective Function

2014 ◽  
Vol 519-520 ◽  
pp. 811-815
Author(s):  
Xiao Hong Qiu ◽  
Yong Bo Tan ◽  
Bo Li
Keyword(s):  
Objective Function ◽  
Detrended Fluctuation Analysis ◽  
Optimization Problem ◽  
Optimization Problems ◽  
Fluctuation Analysis ◽  
Hurst Index ◽  
Test Functions ◽  
Objective Functions ◽  
Fractal Parameters ◽  
Detrended Fluctuation

The fractals of the optimization problems are first discussed. The multi-fractal parameters of the optimal objective function are computed by the Detrended Fluctuation Analysis (DFA) method. The multi-fractal general Hurst Index is related to the difficulty to solve the optimization problem. These features are verified by analyzing the first six test functions proposed on 2005 IEEE Congress on Evolutionary Computation. The results show that the different objective functions have obvious different multifractal and the general Hurst Index can be used to evaluate the difficulty to solve the optimization problem.

scholarly journals Chemical Accuracy in Modeling Halide Ion Hydration from Many-Body Representations

2018 ◽  
Author(s):  
Francesco Paesani ◽  
Pushp Bajaj ◽  
Marc Riera
Keyword(s):  
Computer Simulations ◽  
Water Clusters ◽  
Many Body ◽  
Ion Hydration ◽  
Systematic Analysis ◽  
Ionic Solutions ◽  
Potential Energy Functions ◽  
The Many ◽  
The Individual ◽  
Body Potential

<div> <div> <div> <p>Despite the key role that ionic solutions play in several natural and industrial processes, a unified, molecular-level understanding of how ions affect the structure and dynamics of water across different phases remains elusive. In this context, computer simulations can provide new insights that are difficult, if not impossible, to obtain by other means. However, the predictive power of a computer simulation directly depends on the level of “realism” that is used to represent the underlying molecular interactions. Here, we report a systematic analysis of many-body effects in halide-water clusters and demonstrate that the recently developed MB-nrg full-dimensional many-body potential energy functions achieve high accuracy by quantitatively reproducing the individual terms of the many-body expansion of the interaction energy, thus opening the door to realistic computer simulations of ionic solutions. </p> </div> </div> </div>

Quantitative CBED: an overview

1995 ◽  
Vol 53 ◽  
pp. 130-131
Author(s):  
D.M. Bird
Keyword(s):  
Quantitative Analysis ◽  
High Resolution ◽  
Computer Simulations ◽  
Two Dimensional ◽  
Basic Point ◽  
Beam Equations ◽  
The Many ◽  
Resolution Simulation

In this abstract I will focus on one particular topic which has widespread relevance in the quantitative analysis of CBED; namely the role of computer simulations of CBED patterns. The basic point is that with modern, high-powered workstations we can perform highly accurate and detailed simulations of all types of CBED patterns from crystals in a relatively short amount of time. Most CBED simulations are based on diagonalisation of the many-beam equations, and, for example, on a DEC Alpha workstation one can perform this calculation on a 150 × 150 matrix (sufficient for a highly accurate simulation in many systems) in around 1.8s. A full, two-dimensional CBED pattern with, say, 50 orientations across each disc (which is a high-resolution simulation) can then be calculated in less than 1 hour. How can such simulations be used in a quantitative analysis of CBED patterns? I will address this by looking at some specific examples covering various aspects of quantitative CBED.

scholarly journals Back to Basics: Benchmarking Canonical Evolution Strategies for Playing Atari

2018 ◽  
Author(s):  
Patryk Chrabąszcz ◽  
Ilya Loshchilov ◽  
Frank Hutter
Keyword(s):  
Deep Learning ◽  
Reinforcement Learning ◽  
State Of The Art ◽  
The State ◽  
Evolution Strategies ◽  
Learning Problems ◽  
Local Minima ◽  
Natural Evolution ◽  
The Many ◽  
Made In

Evolution Strategies (ES) have recently been demonstrated to be a viable alternative to reinforcement learning (RL) algorithms on a set of challenging deep learning problems, including Atari games and MuJoCo humanoid locomotion benchmarks. While the ES algorithms in that work belonged to the specialized class of natural evolution strategies (which resemble approximate gradient RL algorithms, such as REINFORCE), we demonstrate that even a very basic canonical ES algorithm can achieve the same or even better performance. This success of a basic ES algorithm suggests that the state-of-the-art can be advanced further by integrating the many advances made in the field of ES in the last decades.We also demonstrate that ES algorithms have very different performance characteristics than traditional RL algorithms: on some games, they learn to exploit the environment and perform much better while on others they can get stuck in suboptimal local minima. Combining their strengths and weaknesses with those of traditional RL algorithms is therefore likely to lead to new advances in the state-of-the-art for solving RL problems.

Topology Optimization for Discovery of Auxetic Origami Structures

2018 ◽  
Author(s):  
Andrew Gillman ◽  
Kazuko Fuchi ◽  
Alexander Cook ◽  
Alexander Pankonien ◽  
Philip R. Buskohl
Keyword(s):  
Topology Optimization ◽  
Modal Analysis ◽  
Design Space ◽  
Element Model ◽  
Mechanical Analysis ◽  
Poisson's Ratio ◽  
Objective Functions ◽  
Geometric Nonlinearities ◽  
Truss Element ◽  
The Many

Origami, as it moves from an art to a scientifically useful technology, enables a rich design space given the numerous bifurcations that exist off the flat state. In this work, we utilize origami as a platform for design of auxetic metamaterials and employ topology optimization for the automated robust discovery of these structures. In particular, the mechanical analysis is performed with an efficient and accurate nonlinear truss element model that captures the geometric nonlinearities associated with origami folding, and modal analysis off the flat state enables access to the many bifurcating branches of folding. Here, objective functions are explored that target a desired in-plane Poisson’s ratio. The Miura-ori fold pattern, a commonly studied flat-foldable pattern, is considered as a verification study for the framework presented.

An Extension of the Embedded Atom Method

1990 ◽  
Vol 193 ◽  
Author(s):  
R. Pasianot ◽  
E. J. Savino ◽  
S. Rao ◽  
D. Farkas
Keyword(s):  
Computer Simulations ◽  
Embedded Atom Method ◽  
Many Body ◽  
Shear Forces ◽  
Embedded Atom ◽  
First Case ◽  
Cauchy Pressure ◽  
The Many ◽  
Three Body ◽  
Embedding Function

For the class of materials in which covalent effects are important, there is still no simple and reliable scheme, adapted to computer simulations, that can handle angle de- pendent forces. Either they are based on the introduction of three body (or higher) [1] interactions, or demand unphysical behavior from the many body functions used [2,3]. In the first case, computer efficiency is considerably low due to the large amounts of calculations required; in the second case a negative curvature of the embedding function must be assumed for materials in which the Cauchy pressure is negative, and this is contrary to the current interpretations of that function.In the present work we derive a method to introduce many body shear forces, suited to computer simulations, which is free from the shortcomings mentioned above.

Sign in / Sign up

Export Citation Format

Share Document