• 품질경영학회지
• /
• 제39권1호
• /
• pp.98-108
• /
• 2011

Clustering is an useful method to classify objects into subsets that have some meaning in the context of a particular problem and has been applied in variety of fields, customer relationship management, data mining, pattern recognition, and biotechnology etc. This paper addresses the unknown K clustering problems and presents a new approach based on a coevolutionary algorithm to solve it. Coevolutionary algorithms are known as very efficient tools to solve the integrated optimization problems with high degree of complexity compared to classical ones. The problem considered in this paper can be divided into two sub-problems; finding the number of clusters and classifying the data into these clusters. To apply to coevolutionary algorithm, the framework of algorithm and genetic elements suitable for the sub-problems are proposed. Also, a neighborhood-based evolutionary strategy is employed to maintain the population diversity. To analyze the proposed algorithm, the experiments are performed with various test-bed problems which are grouped into several classes. The experimental results confirm the effectiveness of the proposed algorithm.

• Industrial Engineering and Management Systems
• /
• 제11권1호
• /
• pp.114-122
• /
• 2012

This paper considers a non-identical parallel machine scheduling problem with sequence and machine dependent setup times. The objective of this problem is to determine the allocation of jobs and the scheduling of each machine to minimize makespan. A mathematical model for optimal solution is derived. An in-depth analysis of the model shows that it is very complicated and difficult to obtain optimal solutions as the problem size becomes large. Therefore, two meta-heuristics, genetic algorithm (GA) and a new population-based evolutionary meta-heuristic called self-evolution algorithm (SEA), are proposed. The performances of the meta-heuristic algorithms are evaluated through compare with optimal solutions using randomly generated several examples.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회/대한산업공학회 2005년도 춘계공동학술대회 발표논문
• /
• pp.614-620
• /
• 2005

Ensemble에서 feature selection은 각 classifier의 학습할 데이터의 변수를 다르게 하여 diversity를 높이며, 이것은 일반적인 성능향상을 가져온다. Feature selection을 할 때 쓰는 방법 중의 하나가 Genetic Algorithm (GA)이며, GA-SVM은 GA를 기본으로 한 wrapper based feature selection mechanism으로 response model과 keystroke dynamics identity verification model을 만들 때 좋은 성능을 보였다. 하지만 population 안의 후보들간의 diversity를 보장해주지 못한다는 단점 때문에 classifier들의 accuracy와 diversity의 균형을 맞추기 위한 heuristic parameter setting이 존재하며 이를 조정해야만 하였다. 우리는 GA-SVM 알고리즘을 바탕으로, population안 후보들의 fitness를 측정할 때 accuracy와 diversity 둘 다 고려하는 fitness function을 도입하여 추가적인 classifier 선택 작업을 제거하면서 성능을 유지시키는 방안을 연구하였으며 결과적으로 알고리즘의 복잡성을 줄이면서도 모델의 성능을 유지시켰다.

• 산업경영시스템학회지
• /
• 제40권1호
• /
• pp.95-104
• /
• 2017

This paper deals with solution methods for discrete and multi-valued optimization problems. The objective function of the problem incorporates noise effects generated in case that fitness evaluation is accomplished by computer based experiments such as Monte Carlo simulation or discrete event simulation. Meta heuristics including Genetic Algorithm (GA) and Discrete Particle Swarm Optimization (DPSO) can be used to solve these simulation based multi-valued optimization problems. In applying these population based meta heuristics to simulation based optimization problem, samples size to estimate the expected fitness value of a solution and population (particle) size in a generation (step) should be carefully determined to obtain reliable solutions. Under realistic environment with restriction on available computation time, there exists trade-off between these values. In this paper, the effects of sample and population sizes are analyzed under well-known multi-modal and multi-dimensional test functions with randomly generated noise effects. From the experimental results, it is shown that the performance of DPSO is superior to that of GA. While appropriate determination of population sizes is more important than sample size in GA, appropriate determination of sample size is more important than particle size in DPSO. Especially in DPSO, the solution quality under increasing sample sizes with steps is inferior to constant or decreasing sample sizes with steps. Furthermore, the performance of DPSO is improved when OCBA (Optimal Computing Budget Allocation) is incorporated in selecting the best particle in each step. In applying OCBA in DPSO, smaller value of incremental sample size is preferred to obtain better solutions.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 2000년도 추계학술대회 및 정기총회
• /
• pp.113-116
• /
• 2000

The purpose of this study is to present a method for determining the k most vital arcs in shortest path problem using an evolutionary algorithm. The problem of finding the k most vital arcs in shortest path problem is to find a set of k arcs whose simultaneous removal from the network causes the greatest increase in the total length of shortest path. The problem determining the k most vital arcs in shortest path problem has known as NP-hard. Therefore, in order to deal with the problem of real world the heuristic algorithm is needed. In this study we propose to the method of finding the k-MVA in shortest path problem using an evolutionary algorithm which known as the most efficient algorithm among heuristics. For this, the expression method of individuals compatible with the characteristics of shortest path problem, the parameter values of constitution gene, size of the initial population, crossover rate and mutation rate etc. are specified and then the effective genetic algorithm will be proposed. The method presented in this study is developed using the library of the evolutionary algorithm framework (EAF) and then the performance of algorithm is analyzed through the computer experiment.

• 경영과학
• /
• 제31권3호
• /
• pp.113-130
• /
• 2014

Cooperative coevolutionary algorithm (CCEA) has proven to be a very powerful means of solving optimization problems through problem decomposition. CCEA implies the use of several populations, each population having the aim of finding a partial solution for a component of the considered problem. Populations evolve separately and they interact only when individuals are evaluated. Interactions are made to obtain complete solutions by combining partial solutions, or collaborators, from each of the populations. In this respect, we can think of various interaction modes. The goal of this research is to develop a CCEA for a supply chain network design (SCND) problem and identify which interaction mode gives the best performance for this problem. We present general design principle of CCEA for the SCND problem, which require several co-evolving populations. We classify these populations into two groups and classify the collaborator selection scheme into two types, the random-based one and the best fitness-based one. By combining both two groups of population and two types of collaborator selection schemes, we consider four possible interaction modes. We also consider two modes of updating populations, the sequential mode and the parallel mode. Therefore, by combining both four possible interaction modes and two modes of updating populations, we investigate seven possible solution algorithms. Experiments for each of these solution algorithms are conducted on a few test problems. The results show that the mode of the best fitness-based collaborator applied to both groups of populations combined with the sequential update mode outperforms the other modes for all the test problems.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 2006년도 추계학술대회
• /
• pp.599-617
• /
• 2006

혼잡한 교통네트워크에서 조사된 통행량으로부터 확률적 사용자 평형을 이루는 통행분포와 통행배정을 동시에 구하기 위한 네트워크 모델과 유전알고리즘을 제안하였다. 확률적 사용자 평형을 이루는 모델은 선형제약을 가진 비선형 목적함수를 최소화하는 문제로 정식화하였다. 네트워크 모델에서는 해의 탐색공간을 줄이고 조사된 통행량을 만족시키기 위해서 흐름보존제약을 활용하였다. 목적함수는 흐름보존, 통행발생량, 통행유입량, 조사통행량 등의 제약을 만족하는 링크통행량과, 경로통행배정을 통하여 구한, 확률적 사용자 평형을 이루는 경로통행량을 만족하는 링크통행량의 차이를 최소화하는 것으로 정식화하였다. 제안된 유전알고리즘에서 유전자는 통행분포, 링크통행량, 여행비용계수 등을 나타내는 벡터로 정의하였다. 각 유전자는 목적함수의 값으로 구한 적합도에 따라 평가되며, 병행단체교차와 돌연변이에 의하여 진화한다.

• Industrial Engineering and Management Systems
• /
• 제11권3호
• /
• pp.288-298
• /
• 2012

Lot sizing and shipment scheduling are two interrelated decisions made by a manufacturing plant and a third-party logistics distribution center. This paper analyzes a dynamic inbound ordering problem and shipment problem with a freight container cost, in which the order size of multiple products and single container type are simultaneously considered. In the problem, each ordered product placed in a period is immediately shipped by some freight containers in the period, and the total freight cost is proportional to the number of containers employed. It is assumed that the load size of each product is equal and backlogging is not allowed. The objective of this study is to simultaneously determine the lot-sizes and the shipment schedule that minimize the total costs, which consist of production cost, inventory holding cost, and freight cost. Because the problem is NP-hard, we propose three meta-heuristic algorithms: a simulated annealing algorithm, a genetic algorithm, and a new population-based evolutionary meta-heuristic called self-evolution algorithm. The performance of the meta-heuristic algorithms is compared with a local search heuristic proposed by the previous paper in terms of the average deviation from the optimal solution in small size problems and the average deviation from the best one among the replications of the meta-heuristic algorithms in large size problems.

• 지구물리
• /
• 제9권4호
• /
• pp.321-331
• /
• 2006

최근 급격한 인구증가와 산업화, 도시화로 포장지역의 증가에 따른 불투수지역의 증가로 유역의 유출 특성의 변화를 유발시키고 있다. 도시화 유역의 효율적인 관리를 위해서는 유역에 대한 정확한 지형인자 및 수문관련 인자들이 추출되어야 함에 따라 본 연구에서는 지리정보체계와 유전자알고리즘의 결합을 통하여 입력정보의 정확성을 향상시키고, 매개변수를 추정하였다. 이러한 목적에 따라 본 연구에서는 전형적인 한국의 도시화하천으로서 본류와 상류로부터 오전천, 당정천 등의 지류를 지니고 있는 안양천을 연구대상으로 선정하여 유출량 해석에 XP-SWMM을 적용하였고, 이의 적용과정을 개선하기 위하여 지리정보체계와 유전자 알고리즘을 적용하였다. XP-SWMM 매개변수들의 민감도 분석을 통하여 도시 유출의 거동특성을 조사하였으며, 이를 바탕으로 매개변수들의 개선규칙을 설정하였고 이러한 규칙 및 사실등을 통하여 유전자 알고리즘을 구성하였다. GIS를 이용하여 지형도로부터 각각의 소유역에 대하여 면적, 경사도, 유역폭 등 수문정보를 얻었고, 토지이용도와 토양도로부터 불투수비, 토지이용상태, 침투능에 대한 정보를 얻었다. 도시유출 모형인 XP-SWMM을 선택하여 모의 후 민감도 분석을 통해 선정된 매개변수에 대하여 보정은 자동보정으로 무작위 탐색법의 일종인 유전자알고리즘(Genetic Algorithm, GA)을 사용하여 매개변수들을 추정하였고, 이의 적용성을 확인하였다.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 1998년도 The Third Asian Fuzzy Systems Symposium
• /
• pp.652-658
• /
• 1998

We introduce the concept of a bi-population scheme for real-coded GAs consisting of an explorer sub-Ga and an exploiter sub-GA. The explorer sub-GA mainly performs global exploration of the search space, and incorporates a restart mechanism to help avoid being trapped at local optima. The exploiter sub-GA performs exploitation of fit local areas of the search space around the neighborhood of the best-so-far solution. Thus the search function of the algorithm is divided. the proposed technique exhibits performance significantly superior to standard GAs on two complex highly multimodal problems.