• The KIPS Transactions:PartB
• /
• v.14B no.7
• /
• pp.541-548
• /
• 2007

Due to the mathematical convergence property, Simulated Annealing (SA) has been one of the most popular optimization algorithms. However, because of its problem of slow convergence in the practical use, many variations of SA like Fast SA (FSA) have been developed for faster convergence. In this paper, we propose and prove that Greedy SA (GSA) also finds the global optimum in probability in the continuous space optimization problems. Because the greedy selection does not allow the cost to become worse, GSA is expected to have faster convergence than the conventional FSA that uses Metropolis selection. In the computer simulation, the proposed method is shown to have as good performance as FSA with Metropolis selection in the viewpoints of the convergence speed and the quality of the found solution. Furthermore, the greedy selection does not concern the cost value itself but uses only dominance of the costs of solutions, which makes GSA invariant to the problem scaling.

• Tunnel and Underground Space
• /
• v.15 no.4 s.57
• /
• pp.275-287
• /
• 2005

At the planning and design stages of a development of underground space or tunneling project, the information regarding ground conditions is very important to enhance economical efficiency and overall safety In general, the information can be expressed using RMR or Q-system and with the geophysical exploration image. RMR or Q-system can provide direct information of rock mass in a local scale for the design scheme. Oppositely, the image of geophysical exploration can provide an exthaustive but indirect information. These two types of the information have inherent uncertainties from various sources and are given in different scales and with their own physical meanings. Recently, RMR has been estimated in unsampled areas based on given data using geostatistical methods like Kriging and conditional simulation. In this study, simulated annealing(SA) is applied to overcome the shortcomings of Kriging methods or conditional simulations just using a primary variable. Using this technique, RMR and the image of geophysical exploration can be integrated to construct the spatial distribution of RM and to evaluate its uncertainty. The SA method was applied to solve an optimization problem with constraints. We have suggested the practical procedure of the SA technique for the uncertainty evaluation of RMR and also demonstrated this technique through an application, where it was used to identify the spatial distribution of RMR and quantify the uncertainty. For a geotechnical application, the objective functions of SA are defined using statistical models of RMR and the correlations between RMR and the reference image. The applicability and validity of this application are examined and then the result of uncertainty evaluation can be used to optimize the tunnel layout.

• 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.

• The KIPS Transactions:PartB
• /
• v.15B no.1
• /
• pp.53-60
• /
• 2008

This paper proposes a new multiobjective optimization method for discriminative training of hidden Markov models (HMMs) used as the recognizer for automatic lipreading. While the conventional Baum-Welch algorithm for training HMMs aims at maximizing the probability of the data of a class from the corresponding HMM, we define a new training criterion composed of two minimization objectives and develop a global optimization method of the criterion based on simulated annealing. The result of a speaker-dependent recognition experiment shows that the proposed method improves performance by the relative error reduction rate of about 8% in comparison to the Baum-Welch algorithm.

• The KIPS Transactions:PartB
• /
• v.8B no.3
• /
• pp.237-242
• /
• 2001

고전적인 GLA 알고리즘과 마찬가지로 Kohonen 학습법은 경도 강하법으로 오차함수의 해에 접근해 나간다. 따라서 KLA의 이러한 문제를 극복하기 위해 모의 담금질법의 일종인 Cauchy 학습법을 응용을 제안한다. 그러나 이 방법은 학습시간이 느리다고 하는 단점이 있다. 본 논문 이 점을 개선시키기 위해 Cauchy 학습법과 Kohonen 학습법을 순차 결합시킨 또 다른 학습법을 제안한다. 그 결과 코시 학습법과 마찬가지로 국부최적 문제를 극복하면서도 삭습시간을 단축할 수 있었다.

• Journal of Korea Multimedia Society
• /
• v.10 no.7
• /
• pp.912-922
• /
• 2007

Evaluation of distance teaming systems needs a method that maintains a consistent level of difficulty for each test. In this paper, we propose a new algorithm for test sheet generation based on genetic algorithm. Unlike the existing methods that difficulty of each test item is assigned by tutors, in the proposed method, that can be adjusted by the result of the previous tests and the average difficulty of test sheet can be consistently reserved. We propose the new genetic operators to prevent duplications of test items in a test sheet and apply the adjusted difficulty of each test item. The result of simulation shows that difficulty of the test sheet generated by proposed method can be more regular than the random method and the simulated annealing method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.10B
• /
• pp.973-980
• /
• 2009

Many studies which have been done for efficient installation and management of wireless sensor networks (WSN) include energy savings, key managements and sensor deployments. Sensor deployment problem is one of the most important and fundamental issues among them in that the topic is directly related with the system cost and performance. In this paper, we suggest a sensor deployment scheme that reduces the system cost of WSN while satisfying the fundamental system requirements of connectivity between sensor nodes and sensing coverage. Using graphical information system(GIS) which contains region-dependent information related with connectivity condition, the initial positions of sensors in the procedure simulated annealing (SA) are determined. The GIS information helps in reducing system cost reduction not only at the initial deployment of SA but also at the final deployment of SA which is shown by computer simulations.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.5 no.1
• /
• pp.33-44
• /
• 2012

A class of recurrent neural networks is proposed and proven to be capable of identifying any discrete-time dynamical system. The applications of the proposed network are addressed in the encoding, identification, and extraction of finite state automata. Simulation results show that the identification of finite state automata using the proposed network, trained by the hybrid greedy simulated annealing with a modified error function in the learning stage, exhibits generally better performance than other conventional identification schemes.