Earticle

This paper presents a new method that can efficiently solve the integrated problem of line balancing and model sequencing in mixed-model U-lines (MMULs). Balancing and sequencing problems are important for an efficient use of MMULs and are tightly related with each other. However, in almost all the existing researches on mixed-model production lines, the two problems have been considered separately. A genetic algorithm for balancing and sequencing in mixed-model U line is proposed. A presentation method and genetic operators are proposed. Extensive experiments are carried out to analyze the performance of the proposed algorithm. The computational results show that the proposed algorithm is promising in solution quality.

최근 많은 실내 사용자들이 증가함에 따라 실내에서의 취약한 통화 품질의 문제가 대두되고 있다. 이러한 문제를 해결하기 위해 셀 소형화 기법이 제시되고 있고, 셀 소형화 기법 중 펨토셀(Femtocell)은 저렴한 설치 비용 대비 고성능의 통화 품질을 제공하기 때문에 문제 해결을 위한 방법으로 많은 연구가 진행되고 있다. 하지만 펨토셀은 사용자에 의해 직접 설치되어 다른 사용자에 대한 간섭을 고려하기 힘들기 때문에, 자기 스스로 주변의 간섭을 고려하여 전력을 설정하여야 한다. 만약 펨토셀 전력이 크게 설정되면 매크로셀 시스템에 간섭으로 작용할 것이며, 반대로 펨토셀 전력이 작게 설정되면 펨토셀 사용자의 수신 성능은 나빠지게 될 것이다. 그러므로 펨토셀 기지국은 매크로셀과 펨토셀 시스템 사이에서 발생하는 trade-off 관계를 잘 고려하여 다른 시스템의 사용자에게는 간섭을 최소화 시키고, 자신의 시스템에 속해 있는 사용자에게는 최대의 신호를 송신하여 좋은 성능을 갖도록 하여야 한다. 본 논문에서는 이러한 trade-off 관계로 인한 문제점을 해결하기 위해 유전자 알고리즘(Genetic Algorithm)을 이용하여 최적의 펨토셀 전력을 설정한다. 또한 가중치 합(Weighted Sum Approach) 기법을 통해 시스템 목적에 따라 다른 가중치를 부여하여 여러 가지 목적에 부합하는 시스템 성능을 갖도록 한다. 컴퓨터 시뮬레이션 수행을 통해 기존의 전력 설정 방식과 제안하는 가중치 합 유전자 알고리즘 기반 펨토셀 전력 설정 방식의 성능을 비교하였으며, 그 결과 제안한 알고리즘의 전력 설정 방식이 사용자에게 더 좋은 SINR(Signal to Noise Interference Ratio)과 많은 채널 용량을 갖는 것을 확인하였다.

Due to the effect of indoor coverage problem, the QoS of the indoor users will be degraded dramatically, with the number of indoor users. The femto cell is a popular solution for such problems. Since the price of the femto base station is usually cheap enough, one can sets up huge number of base stations in a small indoor area to reduce the size of communication cell. In this way, the QoS of the indoor users can be improved significantly. Moreover, the data rate can also be increased. However, how to decide an ideal transmitting power according to the surrounding radio environment is not a trivial problem, that still has not been addressed well. If the transmit power of femto base station is too large, the interference to the macro users will be increased. Conversely, if the transmit power of femto base station is too small; the coverage of femto base station will be reduced. To address this problem, we propose a power configuration method in femto base station using Genetic Algorithm by investigating a new fitness function. Furthermore, we adopt the weighted sum approach to improve the user performance in different modes. The simulation results show that the proposed power configuration method can not only improves the downlink SINR, but also enhance the channel capacity for both the Macro cell systems and Femto cell systems compared with some conventional methods.

본 논문에서는 전기 배전시스템에서 부하단의 실시간 변화에 따른 배전시스템의 안정적인 운용을 가능하게 하기 위해서 유전자 알고리즘 방법을 사용한 방법에 대해서 연구하였다 배전시스템의 안정적인 운용은 각각의 배전 구역에서의 안정성을 향상시킨다는 중요한 장점을 가지고 있다 본 논문에서는 배전계통에서 가장 어려운 것으로 평가되는 접속절차에 대한 접근을 기반의 신뢰성 모델에 기초하여 수행하였다 유전자 알고리즘은 일반적인 생물계에서의 생존을 위한 진화의 과정을 구현한 것으로서 본 논문에서는 개의 노드와 개의 배전영역을 갖는 배전시스템을 대상으로 유전자 알고리즘을 적용한 배전시스템 최적화를 구현하였다

In this paper presents a new approach to evaluate reliability indices of electric distribution systems using genetic Algorithm (GA). The use of reliability evaluation is an important aspect of distribution system planning and operation to adjust the reliability level of each area. In this paper, the reliability model is based on the optimal load transferring problem to minimize load generated load point outage in each sub-section. This approach is one of the most difficult procedures and become combination problems. A new approach using GA was developed for this problem. GA is a general purpose optimization technique based on principles inspired from the biological evolution using metaphors of mechanisms such as natural selection, genetic recombination and survival of the fittest. Test results for the model system with 24 nodes 29 branches are reported in the paper.

Numerous cities developed in the 20th century have a Water Distribution Network(WDN) with decade-old pipes. Aging pipelines can leak with lower the water revenue ratio, thus resulting in waste and economic losses. In this study, new methods for localizing leakage in WDNs are proposed and tested. Two meta-heuristic methods based on Harmony Search(HS) and Genetic Algorithm(GA) are developed to detect leakage of WDNs through an evaluation of emitter coefficients. The old town of G-Town in South Korea has a 50-year-old WDN and a low water revenue ratio. High quality field measurements in G-Town allowed detailed testing of optimization methods based on emitter coefficients for leakage detection. As a result, both GA and HS yielded comparable outputs. The HS method performed with slightly better accuracy in minimizing the objective function than GA after about 1,000 iterations. Leakage detection becomes fairly accurate after approximately 30 minutes of optimization and pipe networks with more than 100 pipes and 50 nodes require more iterations and calculation time.

하플로타입 페이징은 입력으로 주어진 SNP 매트릭스로부터 하플로타입을 결정하는 문제로, 본 논문에서 다루는 PEATH를 포함하여 최근까지도 다양한 하플로타입 페이징 방법들이 제시되고 있다. PEATH는 유전 알고리즘과 토글링(toggling)이라 불리는 휴리스틱을 이용한 하플로타입 페이징 알고리즘으로 많은 계산을 필요로 한다. 본 논 문에서는 PEATH의 수행 시간을 단축시키기 위한 효율적인 구현 방법을 제시하고 실험을 통해 성능을 분석한다. PEATH의 토글링 단계에서는 많은 적합도 계산이 필요한데, 본 연구에서는 계산되는 적합도의 관계를 분석하여, 이전의 계산 결과를 이용하여 적합도를 효율적으로 계산하는 방법을 제안한다. 실험 결과, 제안하는 구현 방법은 PEATH의 수행 시간을 약 12배 개선시켰다.

Haplotype phasing is a problem of, given an SNP matrix as an input, determining its haplotype. Diverse haplotype phasing methods have been proposed. The PEATH is a haplotype phasing algorithm using genetic algorithms and toggling heuristic which require a lot of computation. In this paper we propose an efficient implementation of the PEATH for improving the running time and analyze the performance experimentally. The toggling of the PEATH requires a lot of fitness computations. We analyze the relationship between these fitness computations and propose an efficient implementation of the fitness computations which make use of the previous computation results. The experimental results show that the proposed implementation enhances the running time of the PEATH by about 12 times.

인터넷을 통한 상품 구매 형태의 변화는 고객정보로서 많은 가치를 가지게 되었다. 그리고 많은 새로운 방법을 통하 여 고객의 선호도에 맞는 서비스나 물건 등을 추천하는 추천시스템을 개발하게 되었다. 데이터베이스에 축적된 고객 의 자료를 분석하여 고객관계관리(Customer Relationship Management)가 중요한 마케팅 수단이 되었고, 분 석된 자료를 통하여 고객에게 새로운 서비스나 제품을 추천해주는 추천시스템(Recommendation System)이 마 케팅의 중요한 수단으로 사용되고 있다. 추천시스템에는 새로운 알고리즘들을 접목하여 정확도를 높이는 연구가 진 행 되어오고 있다. 본 논문은 Grouplens의 영화 평가 자료를 이용하였으며, 유전알고리즘을 통하여 얻어진 특징 (feature)들을 기본데이터로 사용하여 추천시스템에 적용하였으며, 추천시스템내의 신경망의 여러 변수들 중 특정 변수값에서 정확도가 높게 나타나는 것을 확인하였다.

Changes in the form of product purchase through the Internet had the value as customer information . Recommendation systems were developed to meet the preferences of customers to recommend such as services or goods through the many new ways. CRM(Customer Relationship Management) is important means of marketing to analyze the data stored in the data server of the customer based. Which is recommended for new services or products to customers through a recommendation system using analyzed data. Marketing has been used as an important tool. Research for recommendation system has been conducted to increase the accuracy by combining the new algorithm. We use the movie rating data, Grouplens. This paper is the feature of the rating data, obtained by genetic algorithm, which are applied to a customer recommendation system. Neural network’s parameters was confirmed high accuracy with specific value in recommendation system.

Testing the effectiveness of quantization matrices in (image) compression is difficult, because each image is unique and requires individualization techniques to optimize the compression ratio. However, there is usually no time to perform these kinds of techniques each time an image is compressed, and thus more general (but less effective) quantization matrices are often used. Although these matrices are standardized, there is still room for improvement. This paper proposes the usage of a popular machine learning technique, genetic algorithms, to actually perform this improvement. Although there might be some generalization issues, we believe it can be partially overcome if the right training set is chosen.

Wireless sensor network (WSN) is a distributed self-organizing network which contains a large number of tiny multi-functional sensor nodes. The network life time is an important issue in WSN because every sensor node has a constraint on electric supply. In this paper, an energy consumption model is described and a GA-based algorithm will be used to optimize the energy consumption by analyzing the working model of sensor nodes. The model will provide an effective reference of working pattern for WSN. This algorithm is evaluated through analysis and simulations.

본 연구에서는 제품의 공급지, 유통센터 및 수요지의 3 계층으로 이루어진 공급사슬 상에서 각 공급지와 각 유통센터간의 1단계 수송비용, 각 유통센터와 각수요지 사이의 2단계 수송비용 및 유통센터의 운영비용의 합을 최소화할 수 있는 제품의 수송계획 문제를 대상으로 한다. 본 연구에서는 유전 알고리즘의 선행 연구를 바탕으로 이 문제에 대하여 우수한 해를 생성할 수 있는 협력적 공진화 알고리즘을 다음과 같이 설계한다. 먼저 이 문제를 2개의 부분문제로 분할하고, 각 부분문제에 대하여 우선순위 기반의 표현을 사용한 염색체 개체들로써 모집단을 구성한다. 그리고 두 모집단내 각 개체에 대한 적합도의 평가 원리를 소개하고, 사용할 적합도 함수를 설정하고, 각 개체에 대한 적합도 평가 알고리즘을 제시한다. 마지막으로 두 모집단의 세대교체에 사용될 선택, 교배 및 돌연변이 연산자를 선정한다. 이러한 설계를 기반으로 만들어진 협력적 공진화 알고리즘을 기존의 유전 알고리즘과 그 성능을 비교해 보기 위하여 여러 개의 테스트 문제에 대하여 반복 실험한다. 그 결과, 협력적 공진화 알고리즘은 문제의 크기가 커짐에 따라 기존의 유전 알고리즘에 비해 해의 평균적인 품질이나 해의 변동성 면에서 상대적으로 더 우수한 해를 얻을 수 있었다. 본 연구에서 제시한 협력적 공진화 알고리즘 기반의 해법은 현실적인 다단계 수송문제에도 확장 적용될 수 있다.

We consider a two stage transportation problem in which the objective is to minimize the total cost including shipping cost from plants to distribution centers, shipping cost from distribution centers to customers, and the opening costs of distribution centers in a three tiers of supply chain consisting of plants, distribution centers and customers. We design a cooperative coevolutionary algorithm to solve the problem as follows. First, the problem is decomposed into two different subproblems. In each subproblem we construct a population of chromosome individuals, each of which is denoting a priority and is represented as a permutation. Then we introduce the evaluation principle for a given individual, set the evaluation function, and suggest the evaluation algorithm. Finally we set selection operator, crossover operator, and mutation operator, each of which is used to generate individuals to be included in the next generation. An experiment study is carried out to compare the performance of our cooperative coevolutionary algorithm with that of the genetic algorithm from previous study. In this experiment we obtained the result that in general our coevolutionary algorithm generates better solution than genetic algorithm as the problem size get larger, in term of both the average quality of solutions and the variance of solutions.

Increasingly, water loss via leakage is acknowledged as one of the main challenges facing WDS (water distribution system) operations. The leakage is losses of an economy, an energy and a water resource. The pressure management is known as one most cost-benefit method in the currently available methods of leakage reduction and prevention. In this paper, an optimization model of water pressure in WDS was developed which was considered a PRV (pressure reduction valve) as a way of pressure management. Xu-Jowitt (1990) model was used to consider a pressure-dependent leakage, GA (genetic algorithm) was used to search a optimal location and numbers of PRV and LP (linear programming) was used to solve a valve control. The model results was compared to EPANET's simulation results for an example network to verify a robustness.

Genetic algorithm based predictor for lossless image compression is propsed. We describe a genetic algorithm to learn predictive model for lossless image compression. The error image can be further compressed using entropy coding such as Huffman coding or arithmetic coding. We show that the proposed algorithm can be feasible to lossless image compression algorithm.

The major task for cognitive radio (CR) systems is to sense the spectrum holes reliably and make the secondary user share the spectrum efficiently and effectively with the primary users. The Cognitive Radio cannot sense the primary users correctly because of the shadowing and fading environment in the multipath channel. To overcome this sensing problem collaboration between the secondary users has to be established. In collaborative sensing, a fusion centre will receive information from each secondary user and a decision will be made by fusion rule. Genetic algorithm is used to select secondary users having higher SNRs. MATLAB is used to carry out the simulation to obtain the optimum number of users can involved in the collaboration. HDL is implemented on Genetic algorithm to detect the optimal secondary users. Simulation and synthesis report is given in this paper.

Owing to the capability to provide a wide variety of intelligent behaviors, cognitive radio (CR) has become a promising technology to improve spectrum utilization efficiently. One of the popular techniques which adapt CR concept to multi-carrier systems is known as tone injection scheme. This scheme is a sort of peak-to-average-power ratio (PAPR) reduction methods deployable to multi-carrier systems such as orthogonal frequency division multiplexing (OFDM). However, a conventional tone injection scheme might increase averaged transmit power attributed to expanding the size of constellation on purpose to get optimal PAPR reduction. Based on a weighted-sum genetic algorithm to resolve multi-objective optimization problem (MOOP), the modified tone injection scheme exploits the agility of CR technology to rapidly adapt operating parameters in order to fulfill PAPR reduction as well as mitigation of power increase optimally. The simulation results verify that the proposed scheme is flexible because it could not only control the performance of PAPR reduction, but also alleviate power increase by steering weight values at the expense of relatively low complexity comparing with other conventional method.

Selected mapping (SLM) is a popular peak-to-average power ratio (PAPR) reduction technique suitable for use in orthogonal frequency division multiplexing (OFDM) systems as it achieves good PAPR reduction performance without signal distortion. However, SLM requires a bank of inverse fast Fourier transforms (IFFTs) to produce candidate signals, resulting in high computational complexity. In this paper, we introduce a novel SLM technique based on conversion matrices (CM) and a genetic algorithm (GA) that requires only one IFFT module. Simulation results indicate that the proposed method obtains desirable PAPR reduction performance with low computational complexity.