• Proceedings of the Korea Multimedia Society Conference
• /
• 2001.11a
• /
• pp.896-899
• /
• 2001

본 논문에서는 외판원 문제를 분산 시스템 환경에서, 다중 에이전트를 이용해 수법시간을 단축시키고, 더욱 우수한 근접해를 구할 수 있는 분산 유전 알고리즘을 개발하였다. 다중 후보해를 이용한 분산 유전 알고리즘을 수행할 때, 고려해야 할 중요한 요소는 후보해들 간의 개체들을 어떤 노드의 후보해 개체와 교환할 것인가와 어떤 개체들을 선택해서, 얼마만큼의 개체를 이동시킨 것인가가 중요하게 고려독어야 한다. 따라서 본 논문에서는 교환해야 할 개체의 크기를 동적으로 윈도우 크기를 변경하면서 교환하는 방법을 개발하였고, 교환할 개체들의 위치를 결정하는 새로운 유전 이동 정책 2가지 방법을 개발하고 실험하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 1998.10c
• /
• pp.470-472
• /
• 1998

본 논문에서는 노이즈와 블러링에 의해 오염된 영상의 비 지도 분할 방법을 제안한다. 본 논문에서는 Markov random field (MRF) model을 사용하는데, 이것은 오염된 여상에 처리하는데 효율적이다. MRF는 연산적으로 복잡하기 때문에 이를 해결하기 위해서 효율적이라는 것과 교통량 측정과 같은 영상 처리에 응용 가능함을 보여준다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2004.05a
• /
• pp.851-854
• /
• 2004

분산 유전 알고리즘은 외판원 문제를 해결하는데 효율적이고 적용하기 쉬운 알고리즘이다. 그러나 다중 후보해를 가진 분산 유전 알고리즘을 수행할 때, 효율성과 정확성에 영향을 주는 많은 요소들이 고려되어야 한다. 후보해의 크기를 얼마로 할 것인지, 이주의 비율 및 횟수는 어떻게 할 것인지와 그리고 어떤 개체들을 선택해서, 어떤 후보해 개체와 교환할 것인가가 중요하게 고려되어야 한다. 따라서 본 논문에서는 이주해야 할 개체의 크기를 동적으로 변경하면서 교환하는 방법과 또한 개체들이 이주되어야 할 위치를 결정하는 이주 정책을 개발하고 실험하였다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.18 no.5
• /
• pp.712-717
• /
• 2008

In swarm robot systems, each robot must act by itself according to the its states and environments, and if necessary, must cooperate with other robots in order to carry out a given task. Therefore it is essential that each robot has both learning and evolution ability to adapt the dynamic environments. In this paper, reinforcement learning method with SVM based on structural risk minimization and distributed genetic algorithms is proposed for behavior learning and evolution of collective autonomous mobile robots. By distributed genetic algorithm exchanging the chromosome acquired under different environments by communication each robot can improve its behavior ability. Specially, in order to improve the performance of evolution, selective crossover using the characteristic of reinforcement learning that basis of SVM is adopted in this paper.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.19 no.2
• /
• pp.279-284
• /
• 2009

In swarm robot systems, each robot must behaves by itself according to the its states and environments, and if necessary, must cooperates with other robots in order to carry out a given task. Therefore it is essential that each robot has both learning and evolution ability to adapt the dynamic environments. In this paper, reinforcement learning method using many SVM based on structural risk minimization and distributed genetic algorithms is proposed for behavior learning and evolution of collective autonomous mobile robots. By distributed genetic algorithm exchanging the chromosome acquired under different environments by communication each robot can improve its behavior ability. Specially, in order to improve the performance of evolution, selective crossover using the characteristic of reinforcement learning that basis of Cascade SVM is adopted in this paper.

• Journal of Wetlands Research
• /
• v.7 no.4
• /
• pp.97-107
• /
• 2005

As the distributed model is developed and widely used, the accuracy of a rainfall measurement and more dense rainfall observation network are required for the reflection of various spatial properties. However, in reality, it is not easy to get the accurate data from dense network. Generally, we could not have the proper rainfall gages in space and even we have proper network for rainfall gages it is not easy to reflect the variations of rainfall in space and time. Often, we do also have missing rainfall data at the rainfall gage stations due to various reasons. We estimate the distribution of mean areal rainfall data from the point rainfalls. So, in the aspect of continuous rainfall property in time, we should fill the missing rainfall data then we can represent the spatial distribution of rainfall data. This study uses the Fuzzy-Genetic algorithm as a interpolation method for filling the missing rainfall data. We compare the Fuzzy-Genetic algorithm with arithmetic average method, inverse distance method, normal ratio method, and ratio of distance and elevation method which are widely used previously. As the results, the previous methods showed the accuracy of 70 to 80 % but the Fuzzy-Genetic algorithm showed that of 90 %. Especially, from the sensitivity analysis, we suggest the values of power in the equation for filling the missing data according to the distance and elevation.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.3
• /
• pp.45-51
• /
• 2011

This paper proposes a method to accelerate the evolution speed of individuals through hybrid reproduction of monogenesis and gamogenesis. Monogenesis as a reproduction method that bacteria or monad without sexual distinction divide into two individuals has an advantage for local search and gamogenesis as a reproduction method that individuals with sexual distinction mate and breed the offsprings has an advantages for keeping the diversity of individuals. These properties can be properly used for improvement of evolution speed of individuals in genetic algorithms. In this paper, we made relatively good individuals among selected parents to do monogenesis for local search and forced relatively bad individuals among selected parents to do gamogenesis for global search by increasing the diversity of chromosomes. The mutation probability for monogenesis was set to a lower value than that of original genetic algorithm for local search and the mutation probability for gamogenesis was set to a higher value than that of original genetic algorithm for global search. Experimental results with four function optimization problems showed that the performances of three functions were very good, but the performances of fourth function with distributed global optima were not good. This was because distributed global optima prevented individuals from steady evolution.

• Structural Engineering and Mechanics
• /
• v.67 no.4
• /
• pp.359-368
• /
• 2018

In this paper, a method is presented to identify the physical and modal parameters of multistory shear building based on substructural technique using block pulse generalized operational matrix and genetic algorithm. The substructure approach divides a complete structure into several substructures in order to significantly reduce the number of unknown parameters for each substructure so that identification processes can be independently conducted on each substructure. Block pulse functions are set of orthogonal functions that have been used in recent years as useful tools in signal characterization. Assuming that the input-outputs data of the system are known, their original BP coefficients can be calculated using numerical method. By using generalized BP operational matrices, substructural dynamic vibration equations can be converted into algebraic equations and based on BP coefficient for each story can be estimated. A cost function can be defined for each story based on original and estimated BP coefficients and physical parameters such as mass, stiffness and damping can be obtained by minimizing cost functions with genetic algorithm. Then, the modal parameters can be computed based on physical parameters. This method does not require that all floors are equipped with sensor simultaneously. To prove the validity, numerical simulation of a shear building excited by two different normally distributed random signals is presented. To evaluate the noise effect, measurement random white noise is added to the noise-free structural responses. The results reveal the proposed method can be beneficial in structural identification with less computational expenses and high accuracy.

• Journal of the Korean Data and Information Science Society
• /
• v.26 no.6
• /
• pp.1479-1494
• /
• 2015

Since the Markowitz's mean-variance framework for portfolio analysis, the topic of portfolio optimization has been an important topic in finance. Traditional approaches focus on maximizing the expected return of the portfolio while minimizing its variance, assuming that risky asset returns are normally distributed. The normality assumption however has widely been criticized as actual stock price distributions exhibit much heavier tails as well as asymmetry. To this extent, in this paper we employ the genetic algorithm to find the optimal portfolio under the Value-at-Risk (VaR) constraint, where the tail of risky assets are modeled with the generalized Pareto distribution (GPD), the standard distribution for exceedances in extreme value theory. An empirical study using Korean stock prices shows that the performance of the proposed method is efficient and better than alternative methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.1
• /
• pp.1-20
• /
• 2018

Service composition in the Inter-Cloud raises new challenges that are caused by the different Quality of Service (QoS) requirements of the users, which are served by different geo-distributed Cloud providers. This paper aims to explore how to select and compose such services while considering how to reach high efficiency on cost and response time, low network latency, and high reliability across multiple Cloud providers. A new hybrid multi-objective evolutionary algorithm to perform the above task called LS-NSGA-II-DE is proposed, in which the differential evolution (DE) algorithm uses the adaptive mutation operator and crossover operator to replace the those of the Non-dominated Sorting Genetic Algorithm-II (NSGA-II) to get the better convergence and diversity. At the same time, a Local Search (LS) method is performed for the Non-dominated solution set F{1} in each generation to improve the distribution of the F{1}. The simulation results show that our proposed algorithm performs well in terms of the solution distribution and convergence, and in addition, the optimality ability and scalability are better compared with those of the other algorithms.