• Proceedings of the KIEE Conference
• /
• 1997.07b
• /
• pp.446-448
• /
• 1997

The aim of this paper is to introduce a method designing the optimal model following boiler-turbine control system using genetic algorithms. This boiler-turbine control system is designed by applying genetic algorithms with reference model to the optimal determination of weighting matrices Q, R that are given by LQ regulator problem. These weighting matrices are optimized simultaneously in the search domain selected adequately. The effectiveness of this boiler-turbine control system is verified by computer simulation.

• Proceedings of the KIEE Conference
• /
• 1997.07b
• /
• pp.587-589
• /
• 1997

In this paper we design the model following ${\mu}$-synthesis control system for optimal fuel-injection of diesel engine using genetic algorithms. To do this, we give gain and dynamics parameters to the weighting functions and apply genetic algorithms with reference model to the optimal determination of weighting functions that are given by D-K iteration method which can design ${\mu}$-synthesis controller in the state space. These weighting functions are optimized simultaneously in the search domain selected adequately. The effectiveness of this ${\mu}$-synthesis control system for fuel-injection is verified by computer simulation.

• Proceedings of the IEEK Conference
• /
• summer
• /
• pp.400-404
• /
• 2004

There are many methods to find solutions for Integer Programming problems (IPs) such as the Branch-Bound philosophy or the Cutting Plane algorithm. However, most of them have a problem that is the explosion of sets in the computing process. In addition, GA is known as a heuristic search algorithm for solutions of optimization problems. It is started from a random initial guess solution and attempting to find one that is the best under some criteria and conditions. The paper will study an artificial intelligent method to solve IPs by using Genetic Algorithms (GAs). The original solution of this was presented in the papers of Fabricio Olivetti de Francaand and Kimmo Nieminen [2003]. However, both have several limitations which causes could be operations in GAs. The paper proposes a method to upgrade these operations and computational results are also shown to support these upgrades.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.03a
• /
• pp.81-84
• /
• 1998

Genetic Algorithms (GAs) are stochastic algorithms whose search methods model some natural phenomena. The procedure of GAs may be divided into two sub-procedures : Operation and Selection. Chromosomes can produce new offspring by means of operation, and the fitter chromosomes can produce more offspring than the less fit ones by means of selection. However, operation which is executed randomly and has some limits to its execution can not guarantee to produce fitter chromosomes. Thus, we propose a method which gives a directional information to the genetic operator by reinforcement learning. It can be achived by using neural networks to apply reinforcement learning to the genetic operator. We use the amount of fitness change which can be considered as reinforcement signal to calcualte the error terms for the output units. Then the weights are updated using backpropagtion algorithm. The performance improvement of GAs using reinforcement learning can be measured by applying the pr posed method to GA-hard problem.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.52 no.2
• /
• pp.114-120
• /
• 2003

According to the use of distribution automation systems, the function to find or to search a fault phase is necessary for automatic switches in a distribution substation. In this paper, two algorithms are developed to fine the fault circuit and the fault phase for the automatic switches in substation with ungrounded power system. One is the fault circuit searching method using the zero sequence voltage at the bus and zero sequence current of circuit current and the other is to find the fault phase using the line voltage and zero sequence current. The developed algorithms are tested in the case study simulations. An ungrounded power system is modeled by EMTP as a case study system. The developed algorithms are tested in the case study simulations and each shows correct results.

• Computers and Concrete
• /
• v.9 no.4
• /
• pp.293-310
• /
• 2012

This paper discusses a new constitutive model called the high-rate brittle microplane (HRBM) model and also presents the details of a new software package called the Virtual Materials Laboratory (VML). The VML software package was developed to address the challenges of fitting complex material models such as the HRBM model to material property test data and to study the behavior of those models under a wide variety of stress- and strain-paths. VML employs Continuous Evolutionary Algorithms (CEA) in conjunction with gradient search methods to create automatic fitting algorithms to determine constitutive model parameters. The VML code is used to fit the new HRBM model to a well-characterized conventional strength concrete called WES5000. Finally, the ability of the new HRBM model to provide high-fidelity simulations of material property experiments is demonstrated by comparing HRBM simulations to laboratory material property data.

• Structural Engineering and Mechanics
• /
• v.52 no.3
• /
• pp.445-468
• /
• 2014

In this paper a new particle swarm ray optimization algorithm is proposed for truss shape and size optimization with natural frequency constraints. These problems are believed to represent nonlinear and non-convex search spaces with several local optima and therefore are suitable for examining the capabilities of new algorithms. The proposed algorithm can be viewed as a hybridization of Particle Swarm Optimization (PSO) and the recently proposed Ray Optimization (RO) algorithms. In fact the exploration capabilities of the PSO are tried to be promoted using some concepts of the RO. Five numerical examples are examined in order to inspect the viability of the proposed algorithm. The results are compared with those of the PSO and some other existing algorithms. It is shown that the proposed algorithm obtains lighter structures in comparison to other methods most of the time. As will be discussed, the algorithm's performance can be attributed to its appropriate exploration/exploitation balance.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.21 no.45
• /
• pp.47-56
• /
• 1998

This paper describes a moving path generation method for the Autonomous vehicles (AV) to search for paths in an unknown environment by using fixed obstacle information. Algorithms for the AV which were recently proposed have some problems, so it was difficult to utilize these algorithms in the real world. The purpose of this research is to examine the applicability of real-time control and efficient improvement by reducing calculation iterations. In the network which is constructed by the cell-decomposition method, a gate is installed in each cell. By verifying the possibility of gate pass-over, the number of cells which should be considered to find the solution can be reduce. Therefore, algorithm iterations can be dramatically improved. In this paper we have proven that path-generated algorithms are efficient by using simulation.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.3
• /
• pp.219-227
• /
• 2002

The post-clustering algorithms, which cluster the result of Web search engine, have some different requirements from conventional clustering algorithms. In this paper, we propose the new post-clustering algorithm satisfying those of requirements as many as possible. The proposed fuzzy Concept ART is the form of combining the concept vector having several advantages in document clustering with fuzzy ART known as real time clustering algorithms on the basis of fuzzy set theory. Moreover we show that it can be applicable to general-purpose clustering as well as post clustering.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.11 no.4
• /
• pp.69-88
• /
• 2007

In this paper we propose new large-update primal-dual interior point algorithms for $P_{\neq}({\kappa})$ linear complementarity problems(LCPs). New search directions and proximity measures are proposed based on a specific class of kernel functions, ${\psi}(t)={\frac{t^{p+1}-1}{p+1}}+{\frac{t^{-q}-1}{q}}$, q>0, $p{\in}[0,\;1]$, which are the generalized form of the ones in [3] and [12]. It is the first to use this class of kernel functions in the complexity analysis of interior point method(IPM) for $P_*({\kappa})$LCPs. We showed that if a strictly feasible starting point is available, then new large-update primal-dual interior point algorithms for $P_*({\kappa})$ LCPs have the best known complexity $O((1+2{\kappa}){\sqrt{2n}}(log2n)log{\frac{n}{\varepsilon}})$ when p=1 and $q=\frac{1}{2}(log2n)-1$.