• 지능정보연구
• /
• 제14권4호
• /
• pp.179-200
• /
• 2008

본 연구에서는 사례기반 추론 기법을 대상으로 효율성과 효과성을 함께 증진시킬 수 있는 속성선정 방법을 개발하였다. 기본적으로, 본 연구에서 개발한 속성선정 방법은 기존에 개발된 단변량 분석 방법과 LVF 알고리즘을 통합하는 것이다. 먼저, 단변량 분석 방법 중 선택효과를 사용하여 전체 속성 중에서 예측력이 우수하다고 판단되는 일부분의 속성들을 추려낸다. 이 속성들로부터 생성해낼 수 있는 모든 가능한 부분집합을 생성해낸 후에, LVF 알고리즘을 이용하여 이 부분집합들이 가지는 불일치 비율을 평가함으로써 최종적으로 속성 부분집합을 선정한다. 본 연구에서 개발한 속성선정 방법을 UCI에서 제공하는 데이터 집합들에 적용하여 성능을 측정한 후, 기존 기법의 성능들과 비교한 결과, 본 연구에서 개발된 속성선정 방법이 선정된 속성의 개수도 만족할만하고 적중률도 향상되어서, 효율성과 효과성 모두의 측면에서 우수함을 보였다.

• 산경연구논집
• /
• 제11권10호
• /
• pp.75-89
• /
• 2020

Purpose: The container terminal is an area that plays an important role in the country's import and export. As the volume of containers increased worldwide, competition between terminals became fiercer, and increasing the productivity of terminals became more important. Re-handling is a serious obstacle that lowers the productivity of terminal. There are two ways to reduce re-handling in the terminal yard. The first method is to load containers in terminal yards using effective carry-in algorithms that reduce re-handling. The second method is to carry out effective remarshalling. In this paper, the performance of various carry-in algorithms and various remarshalling algorithms are reviewed. Next, we try to find the most effective combination of carry-in algorithm and remarshalling algorithm. Research design, data and methodology: In this paper, we analyze the performance of the four carry-in algorithms, AP, MDF, LVF, RP and the four remarshalling algorithms, ASI, ASI+, ASO, ASO+. And after making all the combinations of carry-in algorithms and remarshalling algorithms, we compare their performance to find the best combination. To that end, many experiments are conducted with eight types of 100 bays through simulation. Results: The results of experiments showed that AP was effective among the carry-in algorithms and ASO+ was effective among remarshalling algorithms. In the case of the LVF algorithm, the effect of carrying in was bad, but it was found to be effective in finding remarshalling solution. And we could see that ASI+ and ASO+, algorithms that carry out remarshalling even if they fail to find remarshalling solution, are also more effective than ASI and ASO. And among the combinations of carry-in algorithms and remarshalling algorithms, we could see that the combination of AP algorithm and ASO+ algorithm was the most effective combination. Conclusion: We compared the performance of the carry-in algorithms and the remarshalling algorithms and the performance of their combination. Since the performance of the container yard has a significant effect on the performance of the entire container terminal, it is believed that the results of this experiment will be effective in improving the performance of the container terminal when carrying-in or when remarshalling.