원문정보
Algorithms on layout design for overhead facility
초록
영어
Overhead facility design problem(OFDP) is one of the shortest rectilinear flow network problem(SRFNP)[4]. Genetic algorithm(GA), artificial immune system(AIS), population management genetic algorithm (PM) and greedy randomized adaptive search procedures (GRASP) were introduced to solve OFDP. A path matrix formed individual was designed to represent rectilinear path between each facility. An exchange crossover operator and an exchange mutation operator were introduced for OFDP. Computer programs for each algorithm were constructed to evaluate the performance of algorithms. Computation experiments were performed on the quality of solution and calculations time by using randomly generated test problems. The average object value of PM was the best of among four algorithms. The quality of solutions of AIS for the big sized problem were better than those of GA and GRASP. The solution quality of GRASP was the worst among four algorithms. Experimental results showed that the calculations time of GRASP was faster than any other algorithm. GA and PM had shown similar performance on calculation time and the calculation time of AIS was the worst.
목차
1. 서론
2. 문제 설명
3. 메타휴리스틱 소개
4. 알고리즘 설계
4.1 개체의 설계
4.2 개체의 평가
4.3 돌연변이
4.4 하이퍼 돌연변이
4.5 교차
4.6 지역탐색
4.7 모집단의 상속
4.8 종료조건
4.9 해법별 알고리즘
5. 해법의 비교 분석
5.1 실험 환경
5.2 파라미터 추정
5.3 해법 간 비교 분석
6. 결론
7. 참고문헌