• Journal of the Korean Operations Research and Management Science Society
• /
• v.32 no.4
• /
• pp.1-18
• /
• 2007

In this paper, we develop an adaptive genetic algorithm (aGA). The aGA has an adaptive scheme which can automatically determine the use of local search technique and adaptively regulate the rates of crossover and mutation operations during its search process. For the adaptive scheme, the ratio of degree of dispersion resulting from the various fitness values of the populations at continuous two generations is considered. For the local search technique, an improved iterative hill climbing method is used and incorporated into genetic algorithm (GA) loop. In order to demonstrate the efficiency of the aGA, i) a canonical GA without any adaptive scheme and ii) several conventional aGAs with various adaptive schemes are also presented. These algorithms, including the aGA, are tested and analyzed each other using various test problems. Numerical results by various measures of performance show that the proposed aGA outperforms the conventional algorithms.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.27 no.1
• /
• pp.158-166
• /
• 2003

The paper describes the development and application of advanced evolutionary computing techniques referred to as micro genetic algorithms ($\mu$GA) in the context of engineering design optimization. The basic concept behind $\mu$GA draws from the use of small size of population irrespective of the bit string length in the representation of design variable. Such strategies also demonstrate the faster convergence capability and more savings in computational resource requirements than simple genetic algorithms (SGA). The paper first explores ten-bar truss design problems to see the optimization performance between $\mu$GA and SGA. Subsequently, $\mu$GA is applied to a realistic engineering design problem in the injection molding process optimization.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2000.04a
• /
• pp.347-356
• /
• 2000

This paper compares four models of artificial neural networks (ANN) supported by genetic algorithms the prediction of stock price index. Previous research proposed many hybrid models of ANN and genetic algorithms(GA) in order to train the network, to select the feature subsets, and to optimize the network topologies. Most these studies, however, only used GA to improve a part of architectural factors of ANN. In this paper, GA simultaneously optimized multiple factors of ANN. Experimental results show that GA approach to simultaneous optimization for ANN (SOGANN3) outperforms the other approaches.

• Proceedings of the IEEK Conference
• /
• 2000.06c
• /
• pp.27-30
• /
• 2000

Many human inventions were inspired by nature. The artificial neural network is one example. Another example is Genetic Algorithms(GA). GAs search by simulating evolution, starting from an initial set of solutions or hypotheses, and generating successive "generations" of solutions. This particular branch of AI was inspired by the way living things evolved into more successful organisms in nature. To simulate the process of GA in a computer, we must simulate many times according to varying many GA parameters. This paper describes the implementation of Java Package for efficient applications on Genetic Algorithms, called "JavaGA". The JavaGA used as a application program as well as applet provides graphical user interface of assigning major GA parameters.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.8
• /
• pp.486-494
• /
• 2006

In this paper, we introduce a new solution for the load balancing problem, an important issue in parallel processing. Our heuristic load balancing technique called MGA effectively combines the benefit of both mean-field annealing (MFA) and genetic algorithms (GA). We compare the proposed MGA algorithm with other mapping algorithms (MFA, GA-l, and GA-2). A multiprocessor mapping algorithm simulation has been developed to measure performance improvement ratio of these algorithms. Our experimental results show that our new technique, the composition of heuristic mapping methods improves performance over the conventional ones, in terms of solution quality with a longer run time.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2000.03b
• /
• pp.125-132
• /
• 2000

This study presents the application of genetic algorithms(GA) to the back analysis of tunnels. GA based on the theory of natural evolution, and have been evaluated very effective for their robust performances, particularly for optimizing structure problems. In the back analysis method, the selection of initial value and uncertainty of field measurements influence significantly on the analysis result. GA can improve this problems through a probabilistic approach. Besides, this technique have two other advantages over the back analysis. One is that it is not significantly affected by the form of problems. Another one is that it can consider two known parameter simultaneously. The propriety of this study is verified as the comparison in the same condition of the back analysis(Gens et al, 1987). In this study, it was performed to estimated the geotechnical parameters in the case of weak rock mass at the Kyung Bu Express railway tunnel. GA have been shown for effective application to a geotechnical engineering.

• Computational Structural Engineering
• /
• v.8 no.2
• /
• pp.123-132
• /
• 1995

Genetic Algorithms(GA), which are based on the theory of natural evolution, have been evaluated highly for their robust performances. Traditional GA has mostly used binary code for representing design variable. The binary code GA has many difficulties to solve optimization problems with continuous design variables because of its large computer core memory size, inefficiency of its computing time, and its bad performance on local search. In this paper, a real code GA is proposed for dealing with the above problems. So, new crossover and mutation processes of GA are developed to use continuous design variables directly. The results of read code GA are compared with those of binary code GA for several single and multiple objective optimization problems. As a result of comparisons, it is found that the performance of the real code GA is better than that of the binary code GA, and concluded that the real code GA developed here can be used for the general optimization problem.

• Journal of Institute of Control, Robotics and Systems
• /
• v.17 no.12
• /
• pp.1210-1218
• /
• 2011

GA (Genetic Algorithms) are efficient for searching for global optima but may have some problems such as premature convergence, convergence to local extremum and divergence. These phenomena are related to the evolutionary operators. As population diversity converges to low value, the search ability of a GA decreases and premature convergence or converging to local extremum may occur but population diversity converges to high value, then genetic algorithm may diverge. To guarantee that genetic algorithms converge to the global optima, the genetic operators should be chosen properly. In this paper, we analyze the effects of the selection operator, crossover operator, and mutation operator on convergence properties, and propose the sweeping method of mutation probability and elitist propagation rate to maintain the diversity of the GA's population for getting out of the premature convergence. Results of simulation studies verify the feasibility of using these sweeping operators to avoid premature convergence and convergence to local extrema.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.13 no.3
• /
• pp.337-349
• /
• 2000

Genetic Algorithms(GA) together with simulated annealing are often called methods of last resorts since they can be applicable to any kind of problems, particularly those to which no sophisticated procedures are applicable or feasible. The design of structures is primarily the process of selecting a section for each member from those available in the market, resulting in the problem of combinatorial nature. Therefore it is usual for the design space to include astronomical number of designs making the search in the space often impossible. In this work, Genetic Algorithms and some related technique are introduced and applied to the design of steel frameworks. In problems with a small number of design variables, GA found true global optima. GA also found true optima for the continuous variable test problems and proved their applicability to structural optimization. For those problems of real size, however, it appears to be difficult to expect GA to find optimum or even near optimum designs. The use of G bit improvement added to ordinary GA has shown much better results and draws attention for further research.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.183-186
• /
• 2003

Hybrid genetic algorithms (HGAs) have been studied as various ways. These HGAs usually use both the global search property of genetic algorithm (GA) and the local search one of local search techniques. One of the general types, when constructing HGAs, is to incorporate a local search technique into GA loop, and then the local search technique is repeated as many iteration number as the loop. This paper proposes a new HGA with a conditional local search technique (c-HGA) that does not be repeated as many iteration number as GA loop. For effectiveness of the proposed c-HGA, a conventional HGA and GA are also suggested, and then these algorithms are compared with each other in numerical examples,