• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.52 no.4
• /
• pp.238-245
• /
• 2003

This paper presents a novel method for searching stable solution using DNA coding scheme. Often there are more than one solutions that satisfy the system requirements. These solutions can be viewed as extremes in multimodal function. All extremes are not the same in that some of them are using sensitive to noise or perturbation. This paper address the method that selects a solution that meets the system requirements in terms of output performance and is tolerant to the perspective noise or perturbation. A new method, called a Gradient DNA coding is proposed to achieve such objectives several numerical examples and presented and comparing DNA coding with genetic algorithm is also given.

• The Korean Journal of Applied Statistics
• /
• v.7 no.2
• /
• pp.113-130
• /
• 1994

The algorithm of Oh and Berger (1993) is extended to handle more general cases where the integrand $f(\theta)$ is not only multimodal but also skewed or has some undetected modes, each having curvature not much different from that of the nearest component. It runs Oh and Berger's algorithm in an iterative way, adding a component in each stage to the mixture importance function from previous stage for better approximation between $f(\theta)$ and the importance function.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.32 no.3
• /
• pp.173-181
• /
• 2019

To estimate probabilistic distribution function from experimental data, kernel density estimation(KDE) is mostly used in cases when data is insufficient. The estimated distribution using KDE depends on bandwidth selectors that smoothen or overfit a kernel estimator to experimental data. In this study, various bandwidth selectors such as the Silverman's rule of thumb, rule using adaptive estimates, and oversmoothing rule, were compared for accuracy and conservativeness. For this, statistical simulations were carried out using assumed true models including unimodal and multimodal distributions, and, accuracies and conservativeness of estimating distribution functions were compared according to various data. In addition, it was verified how the estimated distributions using KDE with different bandwidth selectors affect reliability analysis results through simple reliability examples.

• Journal of Advanced Marine Engineering and Technology
• /
• v.21 no.2
• /
• pp.108-116
• /
• 1997

유전알고리즘은 진화원리에서 발견된 몇몇 특징들을 컴퓨터 알고리즘과 결합시켜 복잡한 최적화 문제를 해결하려는 도구로서 1975년 미국의 Holland 교수에 의해 처음으로 개발되었다. 주어진 문제에서 탐색환경이 다변수 또는 다봉(multi-modal)이 되어 대단히 복잡하거나 또는 부분적으로 알려질 경우는, 구배(gradient)에 기초한 재래식 방법을 사용하여 최적화하는 것은 매우 어렵게 되고 경우에 따라서는 불가능할 수도 있다. 이러한 이유로 유전알고리즘과 같은 강인한 탐색법이 요구된다. 유전알고리즘의 장점은 연속성(continuity), 미분가능성(differentiability), 단봉성(unimodality) 등과 같이 탐색공간에 대한 제약으로부터 자유롭다는 것이다. 다시 말하면 목적함수 외 탐색공간에 대한 사전지식을 필요로 하지 않고, 매우 크고 복잡한 공간일지라도 전역해 쪽으로 수렴해 갈수 있다는 것이다. 이러한 특성 때문에 유전알고리즘은 실제 환경에서 많은 복잡한 최적화 문제를 해결하는 방법으로 인정을 받고 있으며, 함수의 최적화, 신경회로망의 학습, 동적시스템의 식별및 제어, 신호처리등 여러 분야에 성공적으로 응용되고 있다. 이러한 중요성에 비해 유전알고리즘에 대한 연구는 국내적으로는 아직 미진한 수준이나 최근 이에 대한 관심이 고조되고 있으며, 또한 그 응용분야도 점점 넓어져 이론 개발과 실질적인 응용에 확산되리라 생각된다. 따라서 본 해설기사는 유전알고리즘의 원리와 응용 사례를 살펴봄으로서 최적화 문제를 해결하려는 독자들에게 조금이나마 도움을 주고자 한다.

• Journal of the Korea Society for Simulation
• /
• v.32 no.1
• /
• pp.23-34
• /
• 2023

The purpose of this paper is to review global stochastic optimization algorithms(GSOA) in case binary response experimentation is used and to compare the performances of them. GSOAs utilise estimator of probability of success $\^p$ instead of population probability of success p, since p is unknown and only known by its estimator which has stochastic characteristics. Hill climbing algorithm algorithm, simple random search, random search with random restart, random optimization, simulated annealing and particle swarm algorithm as a population based algorithm are considered as global stochastic optimization algorithms. For the purpose of comparing the algorithms, two types of test functions(one is simple uni-modal the other is complex multi-modal) are proposed and Monte Carlo simulation study is done to measure the performances of the algorithms. All algorithms show similar performances for simple test function. Less greedy algorithms such as Random optimization with Random Restart and Simulated Annealing, Particle Swarm Optimization(PSO) based on population show much better performances for complex multi-modal function.