유전알고리듬을 이용한 시스템 신뢰도 최적화

다국어 초록 (Multilingual Abstract)

Generally, parallel redundancy is used to improve reliability in many systems. However, redundancy increases system cost, weight, volume, power, etc. Due to limited availability of these resources, the system designer has to maximize reliability subject to various constraints or minimize resources while satisfying the minimum requirement of system reliability.
Genetic Algorithms(GAs), a. search technique. based. on the biological process for natural selection and genetic inheritance，have been shown to be very powerful in a wide variety of applications, particularly in combinatorial optimization problems. The ability of GA to find good solutions efficiently often depends on the encoding, breeding operators, and fitness measures to the given problem.
This study presents GAs to solve optimal redundancy allocation in series systems. To apply the GAs. to this problem，we propose a GA representation，the method for initial population construction，evaluation and genetic operators.
Especially, to improve the performance of GAs, we develop heuristic operators (heuristic crossover, heuristic mutation) using the reliability-resource information of chromosome.
Experiments are carried out to evaluate the performance of the proposed algorithm The performance comparison between the proposed algorithm and a previous method shows that our approach is more efficient.

목차 (Table of Contents)

• Ⅰ. 서론 1
• Ⅱ. 문제의 상황 2
• 1. 시스템 신뢰도 2
• 2. 시스템 신뢰도의 개선 2
• 3. 시스템의 중복설계 3
• 4. 시스템 신뢰도 최적화 문제 4
• Ⅲ. 신뢰도 최적화를 위한 유전알고리듬 5
• 1. 표현 5
• 4.2 초기 모집단 6
• 4.3 적응도 평가 6
• 4.4 선별
• 4.4 유전연산자 7
• 4.4.1 교차 7
• 4.4.2 돌연변이 9
• Ⅳ. 실험 및 분석 10
• Ⅴ. 결론 11